(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Children's Library | Biodiversity Heritage Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
Upload
See other formats

Full text of "Standard Methods of Chemical Analysis: A Manual of Analytical Methods and General Reference for ..."

Google 



This is a digital copy of a book that was preserved for generations on library shelves before it was carefully scanned by Google as part of a project 

to make the world's books discoverable online. 

It has survived long enough for the copyright to expire and the book to enter the public domain. A public domain book is one that was never subject 

to copyright or whose legal copyright term has expired. Whether a book is in the public domain may vary country to country. Public domain books 

are our gateways to the past, representing a wealth of history, culture and knowledge that's often difficult to discover. 

Marks, notations and other maiginalia present in the original volume will appear in this file - a reminder of this book's long journey from the 

publisher to a library and finally to you. 

Usage guidelines 

Google is proud to partner with libraries to digitize public domain materials and make them widely accessible. Public domain books belong to the 
public and we are merely their custodians. Nevertheless, this work is expensive, so in order to keep providing tliis resource, we liave taken steps to 
prevent abuse by commercial parties, including placing technical restrictions on automated querying. 
We also ask that you: 

+ Make non-commercial use of the files We designed Google Book Search for use by individuals, and we request that you use these files for 
personal, non-commercial purposes. 

+ Refrain fivm automated querying Do not send automated queries of any sort to Google's system: If you are conducting research on machine 
translation, optical character recognition or other areas where access to a large amount of text is helpful, please contact us. We encourage the 
use of public domain materials for these purposes and may be able to help. 

+ Maintain attributionTht GoogXt "watermark" you see on each file is essential for in forming people about this project and helping them find 
additional materials through Google Book Search. Please do not remove it. 

+ Keep it legal Whatever your use, remember that you are responsible for ensuring that what you are doing is legal. Do not assume that just 
because we believe a book is in the public domain for users in the United States, that the work is also in the public domain for users in other 
countries. Whether a book is still in copyright varies from country to country, and we can't offer guidance on whether any specific use of 
any specific book is allowed. Please do not assume that a book's appearance in Google Book Search means it can be used in any manner 
anywhere in the world. Copyright infringement liabili^ can be quite severe. 

About Google Book Search 

Google's mission is to organize the world's information and to make it universally accessible and useful. Google Book Search helps readers 
discover the world's books while helping authors and publishers reach new audiences. You can search through the full text of this book on the web 

at |http: //books .google .com/I 



STANDARD METHODS 

OF 

CHEMICAL ANALYSIS 

A MANUAL OF ANALYTICAL METHODS AND 

GENERAL REFERENCE FOR THE 

ANALYTICAL CHEMIST AND FOR 

THE ADVANCED STUDENT 

EDITED BY 

WILFRED W. SCOTT 

Btttareh Cltemut. Oeneral C/uifnieal Company; Formerly 
Chief Chttnitt, Baldwin Loeomotite Worla. AMthor of 
" (iualitatitt Chtmiail Analgtit ; A LaboraUtry Ouiit." 

IN COLLABORATION WITH 

H. A. BAKER D. K. FRENCH R. K. MEADE 

L. C. BARTUN H, A. GARDNER J. C. OLBEN 

F. a. BREYER A.H.GILL R.S.OWENS 

B. H. CLARK F. E. HALE W. L. RAVELL 

W. O. DERHY R. E, HICKMAN J. A. SCHAEPPER 

W. F. DOERFLINOER W. B. HICKB 

142 ILLCSTRJTIoyS AJTI* 3 COLORED PLATES 



SECOND EDITION, REVISED 




NEW YORK 

D. VAN NOSTRAND COMPANY 

25 Pask Place 

1917 



Copyright, 1917, by 
D. VAN NOSTRAND COMPANY 



• • 



• • 



.•• 



« • 






• • 



••• • 
• • • • 
••• • • 

I • • • 
'• • • • • • 



• • 















• • •• 






THIS BOOK IS AFFECTIONATELY DEDICATED 

TO MY FATHER, 



^4059 



PREFACE TO FIRST EDITION 

This book is a compilation of carefully selected methods of 
technical analysis that have proven of practical value to the 
professional chemist. The subjects have been presented with 
sufficient detail to enable one with an elementary knowledge of 
analytical processes to follow the directions; on the other hand, 
lengthy exposition, theoretical dissertation and experimental data 
are purposely avoided, in order to include a large amount of in- 
formation in a compact, accessible form. References to original 
papers are given when deemed advisable. 

For methodical arrangement the material is grouped under 
three major divisions — Part I. Quantitative determination of the 
elements. Part II. Special subjects. Part III. Tables of infor- 
mation. 

In the first division the elements are generally taken up in 
their alphabetical order, each chapter being fairly complete in 
itself, cross-references being given to certain details included 
elsewhere to avoid repetition. For example, the complete di- 
rections for separation of the halogens are given in the chapter 
on chlorine, and references to these details are given in the chap- 
ters dealing with the other members of this group. Occasionally 
it has been deemed advisable to place several related elements 
together in the same chapter. 

Each chapter on the elements is generally arranged according 
to the following outline: 

Physical Properties. Atomic weight; specific gravity; melting- 
point; boiling-point; oxides. 

Detection. Characteristic reactions leading to the recognition 
of the element. 

Estimation. The subject is introduced with such information 
as is useful to the analyst. 

Preparation and Solution of the, Samples. Here directions are 
given for the preparation and decomposition of characteristic 



vi PREFACE 

materials in which the element occurs. Recommendations to the 
best procedures are included to assist the analyst in his choice. 

Separations, This section is devoted to procedures for the 
removal of substances, commonly occurring with the element, 
that may interfere with its estimation. In the absence of such 
substances, or in case methods are to be followed by which a direct 
estimation of the element may be made in the presence of these 
substances, this section on separations may be omitted in the 
course of analysis. Here the discretion of the chemist is necessary, 
and some knowledge of the substance examined essential. 

Methods. The procedures are grouped under gravimetric and 
volumetric methods. Several processes are generally given to 
aflFord the opportunity of selection for particular cases and for 
economical reasons where special reagents may not be available. 

In many of the chapters methods for determining traces of 
the element are given, and the subjects are concluded by typical 
examples of complete analysis of substances containing the elements. 

The titles to the procedures generally give a clue to the processes. 
Names of originators are occasionally retained where common 
usage makes the methods generally known by these. 

Although the combined acid radicals are taken up with the 
elements to which they may be assigned, a chapter is devoted to 
the more important of the acids in their free state, and is placed 
with the other special subjects in the second division of the book. 
Here are found chapters on water, paint, oil, alloys, coal, cement, 
gas, and such subjects as are best classed in sections apart from 
simple substances dealt with in the first portion of the work. 

The last portion of the book is devoted to tables of the more 
important arithmetical operations. These are designed to assist 
the analyst to greater accuracy of calculations, as well as to relieve 
him of needless expenditure of time and energy. 

The material herein included has been carefully selected, an 
effort having been made to obtain the more trustworthy methods 
that will meet the general needs of technical chemists. For guid- 
ance in making certain selections and for information of value to our 
work, we are indebted to the standard works of K. M. Fresenius, 
F. A. Gooch, E. P. Treadwell and W. T. Hall, A. H. Low, J. W. 
Mellor, J. C. Olsen, F. A. Sutton, E. Thorpe and others, as well as 



PREFACE vii 

to the current chemical literature. All unpublished procedures 
appearing in our book have undergone thorough test and have 
proven worthy of a place among standard methods. The col- 
laborators are aware of the limitations of analytical processes and 
will gladly welcome critidsjn of the procedures, and suggestions 
that will enable us to improve the work. 

The editor wishes to acknowledge his indebtedness to those 
cooperating with him in the compilation of this volume. The names 
of these appear in the table of contents, as well as on the title 
pages of their respective chapters. For useful suggestions and 
information of value to this work, or for assistance in reviewing 
manuscript or proof we express our appreciation of Df. F. A. 
Gooch, Mr. W. C. Ferguson, Mr. W. S. Allen, Dr. Allen Rogers, 
Mr. L. E. Barton, Mr. T. T. Gray, Mr. W. G. Derby, Mr. A. W. 
Betts, Mr. N. F. Harriman, Dr. E. Bedtel, Mr. W. F. Doer- 
flinger, Mr. J. M. Cratty, Mr. B. S. Clark and others, mention of 
whom is made in the text. We would make special mention of 
Dr. John C. Olsen for his review of the entire manuscript and 
for many valuable suggestions, which are incorporated in the 
work. We wish to express our high appreciation of Dr. Frank 
E. Hale for his invaluable assistance in reviewing proof and for 
his contributions. 

A list of the majority of publications consulted is given in 
alphabetical order in the appendix of this volume. Reference to 
these authorities will be found throughout the book. 

W. W. Scott. 

New York City, 
January, 1917. 



PREFACE TO SECOND EDITION 

The demand for a second edition of Standard Methods of 
Chemical Analysis within six months of the issue of the first has 
made it impractical to attempt any drastic revision of the work. 
A few errors appearing in the first edition have been corrected and 
some changes made in descriptive portions of certain methods. 
Several useful tables have been added to the data in Part III. 

The Editor trusts that the book wi.l continue to find favor with 
those desiring reliable information in the field of its survey. 

Wilfred W. Scott. 

Grantwood, N. J., 
September, 1917. 



▼ffi 



CONTENTS 

PART I 

ALUMINUM 

Wilfred W. Scott, M. A. 

Research Chemistf General Chemiril Company. Formerly Chief Chemist j Baldwin 
Locomotive Works; Professor of Chemistry, Morningside College. Author of 
** Qualiiative Chemical Analysis; a Laboratory Guide.'* 

Detection — with ammonium hydroxide, with sodium thiosulphate, 3. Estima- 
tion, 3. Preparation and solution of the sample — general procedure for ores, sulphide 
ores, carbonate fusion, bisulphate fusion, extraction of aluminum-bearing ores for 
commercial valuation, metallic aluminum and its alloys, 3-5. Separation of aluminum 
from — silica, iron, phosphoric acid, chromium, manganese, cobalt, nickel, zinc, alkaline 
earths, alkalies, titanium, uranium, glucinum, 5, 6. Gravimetric methods — deter- 
mination of aluminum by hydrolysis with ammonium hydroxide; hydrolysis with 
sodium thiosulphate; precipitation as phosphate; precipitation as aluminum chloride 
10. Volumetric methods — determination of combined alumina by alkali titration; 
free alumina or free acid in aluminum salts — potassium fluoride method, 11-13. Min- 
ute amounts of aluminum by Atack's Alizarine S method, 14. Bauxite analysis, 
14, 15. Aluminum in iron and steel, 16. Analysis of metallic uluminum, standard 
method of the Aluminum Company of America, 17, 18. 

ANTIMONY 

Wilfred W. Scott 

Detection — as sulphide, by hydrolysis, in minerals, traces; distinction between 
antimonous and antimonic salts, 18. Estimation, 19. Preparation and solution of 
the sample — sulphide ores, low grade oxides, speisses, slags, mattes, alloys, hard lead 
rubber goods, 19-21. Separations of antimony from — members of sut>sequent groups; 
from mercury, copper, bismuth, cadmium, lead, arsenic and tin, 21-23. Gravimetric 
methods — determination as antimony trisulphide; as metal by electrolysis, 23, 24. 
Antimony in solder and in alloys with tin and lead, 25. Volumetric methods — de- 
termination with potassium bromate; with potassium iodide; by oxidation with 
iodine; permanganate method; indirect evolution method, 25-28. Determination 
of small amounts of antimony, 28. / 

ARSENIC 

Wilfred W. Scott 

Detection — with hydrogen sulphide, volatility of arsenous chloride, traces; distinc- 
tion between arsenates and arsenites, 30. Estimation, 30. Preparation and solution of 
the sample — pyrites ore, arseno-pyrites, arsenous oxide, arsenic acid, alkali arsenates; 
arsenic in sulphuric acid, in hydrochloric acid, in organic matter; lead arsenate; 

ix 



X CONTENTS 

zinc arsenite; water soluble arsenic in insecticides; arsenic in mispickel; in iron; in 
copper, 30-33. Separations — isolation of arsenic by distillation; separation as sul- 
phide from antimony and tin, etc., 33-36. Gravimetric methods — determination as 
trisulphide; as magnesium pyroarsenate, 36. Volumetric methods — by oxidation 
with standard iodine; by precipitation as silver arsenate, 39, 40. Small amounts of 
arsenic by the Gutzeit method — in sulphuric, hydrochloric, and nitric acids; in iron, 
P3Tites, cinders, bauxite, phosphates, phosphoric acid, salts, baking powder, organic 
matter, canned goods, meats, etc., standard method of the General Chemical Com- 
pany, 40-47. Analysis of commercial arsenic (AsjOi), 47-49. 

BARIUM 

Wilfred W. Scott 

Detection — as barium chromate, with calcium sulphate or strontium sulphate, 
by precipitation as fluosilicate, flame test, spectrum, 50. Preparation and solution 
of the sample — ores, sulphates, sulphides, carbonates; salts soluble in water; organic 
matter; insoluble residue, 50, 51. Separations — the alkaline earths; introductory, 
sources of loss, preliminary tests; separation from members of previous groups; separa- 
tion of the alkaline earths from magnesium and the alkalies by the oxalate and sul- 
phate methods; separation of the alkaline earths from each other, 51-56. Gravi- 
metric methods — determination as chromate; as sulphate, 56-58. Volumetric methods 
— titration of the barium salt solution with dichromate; reduction with ferrous salt 
and titration with permanganate; potassium iodide method; acid titration of the 
carbonate, 59, 60. Analysis of barytes and witherite; commercial valuation of the ores, 
59-61. 

BISMUTH 

Wilfred W. Scott 

Detection — as bismuth oxychloride, and by reduction, 62. Elstimation, 62. Prep- 
aration and solution of the sample — ores, cinders; alloys, bearing metal, lead bullion 
and refined lead, 62, 63. Separations from — members of the ammonium hydroxide, 
sulphide and carbonate groups and from the alkalies; separation from arsenic, anti- 
mony, tin, molybdenum, tellurium and selenium; mercury, lead, copper and cadmium, 
64, 65. Gravimetric methods — by precipitation and determination as the basic 
chloride, BiOCl; as the oxide, BisOj, (a) by precipitation as the basic nitrate, (6) sub- 
carbonate, (c) hydroxide, determination as bismuth sulphide, BiaSs; as metallic bis- 
muth by reduction with potassium cyanide; by deposition of the metal electrolytically, 
65-68. Volumetric methods — by titration of the oxalate with permanganate; cin- 
chonine potassium iodide colorimetric method; bismuth iodide colori metric com- 
parison, 68-70. 

BORON 

Wilfred W. Scott 

Detection — flame test, borax bead and turmeric tests, 71. Estimation, 71. Prep- 
aration and solution of the sample — boric acid in silicates and enamels; boronatro- 
calcite, borocalcite, boracite, calcium borate; borax and boric acid; boric acid in min- 
eral water; in carbonates; in foods — milk, butter, meat, etc., 72, 73. Gravimetric 
methods — distillation of methyl borate and fixation with lime, 74, 75. Volumetric 
methods — titration of boric acid in presence of mannitol or glycerole in evaluation of 
borax or boric acid. Robin's test for traces, 76, 77. 



CONTENTS xi 

BROMINE 

Wilfred W. Scott 

Detection — ^by sflver nitrate, by absorption in carbon tetrachloride or disulphide, 
by magenta test, bromates, 78. Estimation, 79. Preparation and solution of the sam- 
ple — bromides, bromine in organic matter, 79. Separation of bromine from the heavy 
metals, from silver, cyanides, chlorine and iodine, 79, 80. Gravimetric methods — 
precipitation as silver bromides, (1) hydrobromic acid and bromides of the alkaline 
earths and alkalies. (2) treatment in presence of heavy metals, 80. Volumetric 
methods — determination of free bromine with potassium iodide; soluble bromides 
by chlorine method, Volhard's method; traces of bromine, 80-82. Arscnous acid 
method for bromates, 82. Analysis of crude potassium bromide and commercial 
bromine, 82, 83. 

CADMIUM 

Wilfred W. Scott 

Detection — as cadmium sulphide, blowpipe test for, 84. Estimation, 84. Prep- 
aration and solution of the sample — sulphides, treatment in presence of lead, car- 
bonates, alloys, 84, 85. Separations from — ^silica, ammonium sulphide group, alkaline 
earths and alkalies, separation from lead, bismuth, mercury, copper, arsenic, anti- 
mony and tin, SS, 86. Gravimetric methods — determination as sulphate; electro- 
lytic method, 86, 87. Volumetric determination by titration with standard iodine 
solution, 87. 

CALCIUM 

Wilfred W. Scott 

Detection — as oxalate, flame test, spectrum of calcium, 88. Estimation and 
occurrence, 88. Preparation and solution of the sample — limestone, dolomite, mag- 
nesite, cement, lime, g3rpsum. Plaster of Paris, sulphates, silicates, chlorides, nitrates 
and other water soluble salts, sulphides, pyrites ore, 89. Separation of calcium from — 
silica, iron, alumina, copper, nickel, cobalt, manganese, zinc, barium, strontium and 
the alkalies, treatment in presence of phosphates of iron and aluminum, 89, 90. Gravi- 
metric methods — precipitation as calcium oxalate; other methods, 91-92. Volumetric 
method by titration of the oxalate with standard potassium permanganate solution, 93. 

CARBON 

Wilfred W. Scott 

Detection — element, carbon dioxide in carbonates and in gas, free carbonic acid 
in water, distinction between carbonates and bicarbonates, carbon monoxide, 93. 
Estimation, 94. Preparation of the sample — iron, steel and alloys; organic matter; 
carbonates and bicarbonates, 94. Separation of carbon from other substances; separa- 
tion from iron by the cupric potassium chloride method, 94, 95. Gravimetric method — 
combustion furnaces, types of absorption apparatus, general procedure for determin- 
ing carbon by combustion, 95-99, 100. Graphitic carbon in iron and steel, 99, 100. 
Combined carbon, 100, 101. Determination of carbon in organic substances — organic 
matter free of nitrogen, halogens, sulphur and the metals, 101; carbon and hydrogen 



xii CONTENTS 

in nitrogenous substances, 102; organic substances containing halogens, 102; wet 
combustion process, 102, 103. Determination of carbon dioxide in carbonates — 
gravimetric, 103-107. Residual and available carbon dioxide in baking powder, 105- 
106. Determination of carbon dioxide by measuring the gas, 105 (107, reference). 
Determination by loss of weight, 106. Volumetric methods, total carbon by barium 
hydroxide absorption, 107. Direct colorimetric method for determining carbon in 
iron and steel, 108-110. Analysis of graphite, 110. Volumetric determination of 
hydrocyanic acid in soluble cyanides, 110, 111. 

CERIUM AND OTHER RARE EARTHS 

R. Stuart Owens 

Research Chemist^ New York City 

Members of the rare earth group — atomic weights, specific gravities, melting- 
points, oxides. Detection of cerium, lanthanum, praseodymium, neodymium, scan- 
dium, ytterbium and erbium, 112, 113. Estimation — occurrence, 113. Preparation 
and solution of the sample, 114. Separations — rare earths from iron, aluminum, 
thorium, calcium, magnesium; separation of scandium from yttrium, yttrium group, 
separation of praseodymium, neodymium, lanthanum and samarium from each other, 
114, 115. Gravimetric estimations, 115. Volumetric method for the determination 
of cerium. Determination of cerium in Welsbach mantles, colorimetric method, 116.. 
Separation of the rare earth oxalates (outline table), 117. 

CHLORINE 

Wilfred W. Scott 

Research Chemist^ General Chemical Company 

Wm. F. Doerflinger 

Chief Chemist, Perry-Austin Manufacluring Company 

Detection — free chlorine, chlorides by silver nitrate test, free hydrochloric acid, 
detection of chlorine in presence of cyanate, cyanide, thiocyanate, bromide, iodide, 
chlorate. Test for hypochlorite, chlorite, chlorate, perchlorate, 118, 119. Estimation 
— occurrence, 119. Preparation and solution of the sample — ^water-soluble chlorides, 
water- insoluble chlorides, silver chloride, chlorine in rocks, free chlorine, chlorine in 
ores and cinders, 119, 120. Determination of halogens in organic compounds — Carius 
method, 121; lime method, 122; sodium peroxide method, 122. Separations — chlorine 
and the halides from the heavy metals, the halides from silver and silver cyanide; 
separation of the halides from one another, chlorine from iodine, and from bromine 
and iodine, 123, 124. Gravimetric method — determination of chloride by precipita- 
tion as silver salt, 124. Volumetric methods — silver thiocyanate ferric alum method 
of Volhard, silver chromate method of Mohr, volumetric method for determination of 
free chlorine, 125-127. Determination of hypochlorous acid in presence of chlorine, 
127. Gravimetric method for determination of chloric acid, 128. Gravimetric deter- 
mination of perchloric acid, 128. Determination of chlorates and pcrchlorates in 
presence of one another, 129. Determination of hydrochloric, chloric, and per- 
chloric acids in presence of one another, 129. Estimation of chlorine, bromine and 
iodine in presence of each other, 130. Evaluation of bleaching powder, chloride of 
lime, for available chlorine, 130. 



CONTENTS xui 

CHROMnTM 

« 

Wilfred W. Scott 

Detection — test with barium salt, hydrogen peroxide, reducing agents, ether, 
diphenyl carbazide, 132. Estimation, 132. Preparation and solution of the sample — 
general procedures for decomposition of refractory materials, special procedures — ^high 
silica ores, chrome iron ores, iron and steel, 133, 134. Separations — chromium from 
iron and aluminum, 134. Gravimetric methods — precipitation of chromic hydroxide 
and ignition to OsOs, determination as barium chromate, 135, 136. Volumetric 
methods — ^potassiimi iodide method, procedure by reduction with ferrous salts, 137-138. 
Determination of small amounts of chromium, 138. 

COBALT 

W. L. Savell, Ph.D. 

Research Chemist, Doloro Mining and Reduction Company 

Detection — general procedure, potassium sulphocyanate, potassium nitrite, 140. 
Estimation, 141. Preparation and solution of the sample — general procedure for 
ores, cobalt oxides, metallic cobalt, nickel and cobalt alloys, 141. Separations — 
anunonium sulphide group from hydrogen sulphide, anmionium sulphide group from 
the alkalies, cobalt and nickel from manganese, cobalt from nickel, cobalt from zinc, 
142. Gravimetric methods — precipitation of cobalt with potassium nitrite, nitroso- 
betarnaphthol method, electrol3rtic deposition of metallic cobalt, 143-145. Deter- 
mination of cobalt in cobalt oxide, 145. Cobalt in the commercial metal and in ferro- 
cobalt, 146. Cobalt in metallic nickel, 146. Cobalt in ores and in enamels, 147. 
Determination of cobalt in steel, 148. 

COPPER 

Wilfred W. Scott 

Research Chemist, General Chemical Company 

Wallace G. Derby 

Research Chemist, Nichols Copper Company 

Detection — general procedure, flame test, wet tests, hydrogen sulphide test, other 
methods, 149. Estimation — occurrence, 149. Preparation and solution of the sam- 
ple — solubilities, decomposition of copper ores, sulphide ores, copper pyrites, copper 
glance, iron pyrites, matte, oxidized ores, oxides, treatment of matte slag, metals, 
iron ores and iron ore briquettes, steel, cast iron and alloy steels, 150-153. Separa- 
tions — precipitation of copper as sulphocyanate, separation of copper by precipitation 
in metallic form by a more positive element, separation from members of the am- 
monium sulphide and subsequent groups; removal of silver, removal of bismuth, lead, 
mercury, arsenic, antimony and tin; separation from cadmium, 153-155. Gravi- 
metric methods — Deposition of metallic copper by electrolysis — introduction. Rapid 
methods — solenoid method of Heath, deposition from nitric acid solution, deposition 
form ammoniacal solution. Slow methods — electrolytic determination of copper in 
blister copper, standard procedure of Nichols Copper Company, large portion and 
small portion methods, traces of copper in the electrolyte, notes and precautions for 
the electrolytic deposition of copper, 155-162. Other methods — determination as 
cuprous sulphocyanate, determination as copper oxide, 162. Volumetric methods 



xiv CONTENTS 

I for determining copper — ^potassiiim iodide method, potassium cyanide method, 

I 163-165. Colorimetric determination of small amoimts of copper — potassium ethyl 

{ xanthate method, ferrocyanide method, ammonia method, hydrogen sulphide method, 

165-167. Determination of impurities in blister copper — ^bismuth, iron, lead, zinc, 
nickel, cobalt, arsenic, antimony, selenium, tellurium, oxygen, sulphiu*, phosphorus, 
167-173. Determination of copper in refined copper, 173. Chlorine in cement 
copper and copper ores, 174. Copper in blue vitriol, 175. Copper and lead deter- 



mination in brass, 175. 



FLUORINE 

Wilfred W. Scott 



Detection — etching test, hanging drop test, black filter paper test, 176-178. Esti- 
mation — occurrence, 178. Preparation and solution of the sample — solubilities, or- 
ganic substances, silicious ores and slags, calcium fluoride, soluble fluorides, hydro- 
fluoric acid, fluorspar, 178, 179. Separations — ^removal of silicic acid from fluorides, 
separation of hydrofluoric and phosphoric acids, separation of hydrofluoric from 
hydrochloric and boric acids, 179, 180. Gravimetric methods — precipitation as 
calcium fluoride, as lead chlorofluoride, 180, 181. Volumetric methods — OfTerman's 
method, ferric chloride method, Steiger and Merwin colorimetric method, 182-186. 
Valuation of fluorspar, standard method of the Faurview Fluorspar and Lead Com- 
pany, 186. Analysis of sodium fluoride, 187. Determination of traces of fluo- 
rine, 188. 

GLUCINUM (BERTLLinM) 

Wilfred W. Scott 

Detection, 189. Estimation — occurrence, 189. Separations — ^removal of silica 
and of the hydrogen sulphide group, separation of glucinum from iron, manganese, 
zirconium, yttrium, aluminum, chromium, iron, 190. Gravimetric determination of 
glucinum, 190, 191. 

GOLD 

Wallace G. Derby, M.S. 

Research Chemist, Nichols Copper Company 

Detection of gold in alloys, test for gold in minerals, benzidine acetate tests, phenyl- 
hydrazine acetate test, 192, 193. Estimation — solubility, 193. Gravimetric methods — 
wet gold assay of minerals, electrolytic method, 194. Volumetric methods — ^per- 
manganate and iodide methods, 195, 196. Colorimetric methods — ^procedures of 
Prister, Cassel, Moir, 197, 198. Preparation of proof gold, 198. 

IODINE 

Wilfred W. Scott 

Detection — element, free iodine characteristics, iodide, iodate, 200. Estimation — 
occurrence, 200. Preparation and solution of the sample, iodides of silver, copper, 
mercury, lead, etc., iodates, free iodine (commercial crystals), iodine or iodides in water, 
organic substances, mineral phosphates, 200, 201. Separations — iodine from heavy 
metals, from bromine or from chlorine, separation from chlorine and bromine — palla- 
dous iodide method, 202, 203. Gravimetric methods — precipitation of silver iodide. 



CONTENTS XV 

detenmnation as palladous iodide, 203. Volumetric methods — ^hydriodic acid and 
iodides by thiosulphate or by arsenite titration, decomposition by ferric salts, decom- 
position with potassium iodate, nitrous acid method of Fresenius, hydrogen peroxide — 
pbosphoric add method, chlorine method of Mohr, Volhard's method, 203-207. 
Determination of iodates, periodates, and iodates with periodates in a mixture, 208, 209. 

moN 

Wilfred W. Scott 

Detection — ferric iron, hydrochloric acid solution, sulphocyanate, ferrocyanide, 
salicylic acid, sodium peroxide tests; distinction between ferrous and ferric salts, 210, 
Estimation — occurrence, 210. Preparation and solution of the sample — solubilities' 
soluble iron salts, sulphide and oxide ores, iron ore briquettes, silicates, iron and steel, 
211, 212. Gravimetric methods — determination as oxide, FeiOa; Cupferron method, 
213, 214. Volumetric methods — general considerations, by oxidation, the iron having 
been reduced — titration with potassium dichromate, potassium permanganate, Jones 
redactor method, 214-221. Stannous chloride method for determining ferric iron, 
221. Colorimetric methods for determining small amounts of iron — with sul- 
phocyanate, with salicylic acid, 222, 223. Technical analysis of iron and steel — 
^th specifications of Baldwin Locomotive Works. Introduction; preparation of 
the sample, combined or carbide carbon-colorimetric method, specifications for 
combined carbon; total carbon by combustion; graphite in iron; manganese by 
persulphate method, lead oxide procedure of Deshey, method by U. S. Bureau of 
Standards — manganese by the Bismuth ate method; specifications for manganese; 
determination of phosphorus by the alkalimetric and molybdate methods; specifica- 
tions for phosphorus; determination of sulphiu* by the evolution method; method by 
the U. S. Bureau of Standards; specifications for sulphur in iron and steel; determi- 
nation of silicon, rapid foundry method, method by the U. S. Bureau of Standards; 
specifications for silicon in iron and steel, 223-232. 

LEAD 

Wilfred W. Scott 

Detection — hydrochloric acid test, hydrogen sulphide test, confirmation of tests, 
233. Estimation — occurrence, 233. Preparation and solution of the sample — sol- 
ubilities, decomposition of ores, minerals of lead, iron pyrites, alloys, 234. Separations 
— ^isolation of lead as sulphate, separation from barium, isolation as lead chloride, 
ammonium acetate extraction of lead from the impure sulphate, 235. Gravimetric 
methods — determination of lead as sulphate, as chromate, as molybdate, electrolytic 
determination as peroxide, Pb02, 236-238. Volumetric methods — ^ferrocyanide method, 
molybdate method, 238-240. Determination of small amounts of lead — gravimetric 
methods — ^by ammonium acetate extraction, by occlusion with iron hydroxide, Seeker- 
Clayton method modified; colorimetric estimation, 241-247. Analysis of metallic 
lead, method of the National Lead Company, modified, 248-252. 

MAGNESIUM 

Wilfred W. Scott 

Detection — general procedure, test with baryta or lime water, 253. Estimation — 
occurrence, 253. Preparation and solution of the sample — solubilities, general pro- 



xvi CONTENTS 

cedure for ores, 254. Separations — removal of members of the hydrogen sulphide 
group, ammonium sulphide group, separation from the alkaline earths, 254. Gravi- 
metric determination by precipitation as ammonium magnesium phosphate, 255. 
Volumetric determination by titration of ammonium magnesium phosphate with stand- 
ard acid, 256. 

MANGANESE 

Wilfred W. Scott 

Detection — general procedure, manganese in soils, minerals, vegetables, etc.* 
borax test, sodium carbonate and nitrate tests, 257. Estimation — occurrence, 257. 
Preparation and solution of the sample — solubilities, decomposition of ores, sulphides, 
slags, iron ores, alloys, manganese bronze, fcrro-titanium alloy, ferro-chromiumi 
metallic chromium, ferro-aluminum, vanadium alloys, molybdenum allo3rs, tungsten 
alloys, silicon alloys, iron, steel and pig iron, 258, 259. Separations — removal of the 
members of the hydrogen sulphide group, separation of manganese from the alkaline 
earths and the alkalies, from nickel and cobalt; basic acetate method for removal of 
iron and aluminum; isolation of manganese as the dioxide, Mn02, 260-262. Gravi- 
metric method — determination as manganese pyrophosphate, 262. Volumetric 
methods — bismuthatc method, procedure of Volhard, ammonium persulphate colori- 
metric method, oxidation of manganese with oxides of lead, 263-268. Analysis of 
Spiegel iron for manganese, 268. 

MERCURY 

Wilfred W. Scott 

Detection — general procedures, 270. Estimation — occurrence, 270. Preparation 
and solution of the sample — solubilities, decomposition of ores, 270, 271. Separations 
— removal of mercury in presence of members of the ammonium sulphide, ammo- 
nium carbonate and soluble groups; separation of mercury from arsenic, antimony, 
tin, lead, bismuth, copper, cadmium, selenium, tellurium, organic substances, 271. 
Gravimetric methods — precipitation as sulphide, determination by electrolysis, HoUo- 
way-Eschka process, 272-274. Volumetric process by Seamon's method, 274. 

MOLYBDENUM 

Wilfred W. Scott 

Detection — general procedure, sodium thiosulphate test, sulphur dioxide test, 
phosphate test, detection in minerals, characteristics of molybdenite, 275. Estima- 
tion — occurrence, 275. Preparation and solution of the sample — solubilities, decom- 
position of ores, steel and iron, 276. Separations — molybdenum from iron in presence 
of large amounts of iron, separation from the alkalies, alkaline earths, lead, copper, 
cadmium, bismuth, vanadium, arsenic, phosphoric acid, titanium and tungsten, 276, 
277. Gravimetric methods — precipitation as lead molybdatc, determination as oxide, 
MoOs after precipitating with mercurous oxide, determination as sulphide, 278, 279. 
Volumetric methods — iodomctric reduction method, estimation by reduction with 
Jones reductor and subsequent i>ermanganate titration, determination of molybdenum 
and vanadium in a mixture of the two, 280-282. 



CONTENTS xvii 

NICKEL 

W. L. Savell, Ph.D. 

Research ChemUtf Doloro Mining and Reduction Company 

Detection — general procedure, dimethylglyoxime test, alpha benzildioxime test, 
283. Estimation, 284. Preparation and solution of the sample, solubilities, general 
procedure for decomposing ores, fusion methods, solution of metallic nickel and its 
alloys, 284, 285. Separations — ammonium sulphide group from hydrogen sulphide 
and from the alkaline earths and alkalies; separation of nickel from cobalt, manganese, 
line, iron, aluminum, chromium, 285, 286. Gravimetric methods — ^precipitation of 
nickel by alpha benzildioxime, precipitation by dimethylglyoxime, electrolytic deposi- 
tion of nickel, nickel in metallic nickel, in cobalt and cobalt oxide, 286-290. Volu- 
metric determination of nickel in alloys, determination in nickel-plating solutions, 290. 

NITROGEN 

Wilfred W. Scott 

Detection — organic nitrogen, nitrogen in gas mixtures. Ammonia, free and com- 
bined, tests for. Nitric acid — ferrous sulphate test, diphenylamine test, copper test, 
phenolsulphonic acid test. Detection of nitrous acid — ^by acetic acid, by potassium 
permanganate, 291, 292. Estimation — occurrence, composition of air; free nitrogen, 
total nitrogen, combined nitrogen, 292, 293. Preparation of the sample — organic 
substances — method in absence of nitrates, method in presence of nitrates; soils — 
available nitrate, ammonium salts, nitrates, nitrites, mixtures of ammonium salts, 
nitrates and nitrites; nitric acid in mixed acid, 293-295. Separations — ammonia, 
nitric acid, removal of nitrous, chromic, hydrobromic and hydriodic acids, 295, 296, 
Procedures for the determination of combined nitrogen — ammonia — gravimetric deter- 
mination of ammonia by precipitation as ammonium platinochloride, 296. Volu- 
metric methods — analysis of aqua ammonia, combined ammonia in ammonium salts; 
analysis of ammoniacal liquor — ammonia, carbon dioxide, hydrochloric acid, hydro- 
gen sulphide, sulphuric acid; determination of traces of ammonia (ref.), 296-299. 
Nitric acid — gravimetric determination by precipitation as nitron nitrate, 299. Volu- 
metric methods — direct estimation of nitrates by reduction to ammonia by the Allen- 
Devarda method, 300. Analysis of nitrate of soda — moisture, insoluble matter, sodium 
sulphate, iron, alumina, lime, magnesia, sodium chloride and carbonates, 303, 304. 
Nitric nitrogen in soU extracts by Vamari-Mitscherlich-Devarda method, 304. De- 
termination of nitrogen of nitrates and nitrites by means of the nitrometer — general 
procedure, DuPont nitrometer method, 305-309. Determination of nitric acid in 
oleum by DuPont nitrometer method, 309. Combined nitric acid (reference to ferrous 
sulphate method), 209. 

PHOSPHORUS 

Wilfred W. Scott 

Detection — element, acids — hypophosphorous, phosphorous, orthophosphoric, 
metaphosphoric, pyrophosphoric, comparative table, 310, 311. Estimation — occur- 
rence, tj^pical ores, 311. Preparation and solution of the sample — iron ores, phosphate 
rocks, minerals, iron and steel, ores with titanium, soluble phosphates, baking powder, 
etc. Precipitation of ammonium phosphomolybdate, 312-314. Gravimetric methods 
^Hlirect weighing of ammonium phosphomolybdate, determination as magnesium 



xviii CONTENTS 

pyrophosphate; direct precipitation of magnesium ammonium phosphate, 314r-316. 
Volumetric methods — by titration with an alkali, by titration of the reduced molyb- 
date with potassium permanganate, 316-318. Phosphate rock analysis — (tentative 
methods of the committee on standard methods), sampling and determination of mois- 
ture, phosphoric acid, iron and aluminum phosphates, iron, 319-323. 

PLATINUM 
Reginald E. Hickman 

Chief Chemistf J. Bishop and Company Platinum Works 

Detection — general characteristics, tests with potassium iodide, hydrogen sulphide, 
ammonium chloride, potassium chloride, ferrous sulphate, stannous chloride, oxalic 
acid, sodium hydroxide with glycerine, formic acid, etc., 324, 325. Estimation — 
characteristic substances containing platinum^ 325. Preparation and solution of the 
sample — solubilities, decomposition of ores, platinum scrap, substances containing 
small amounts of platinum, 325-327. Separations — platinum from gold, from iridium, 
palladium, ruthenium, rhodium, osmium, 327, 328. Gravimetric methods — ^weighing 
as metallic platinum, weighing as a salt, electrolytic method, 328, 329. 

RARER ELEMENTS OF THE ALLIED PLATINUM METALS 

Reginald E. Hickman 

Iridium. — Detection — tests with caustic alkalies, potassium chloride, ammonium 
chloride, hydrogen sulphide, metallic zinc, formic and sulphurous acids, 330. Estima- 
otin. Preparation and solution of the sample, 330. Separations — iridium from plat- 
inum, 331. Gravimetric methods, reduction with zinc, ignition of the ammonium 
salt, insoluble residue, 331, 332. 

Palladium. — Detection — tests with alkalies, ammonia, mercuric cyanide, potassium 
iodide, hydrogen sulphide, etc., 332, 333. Estimation — Preparation and solution 
of the sample. Separations — palladium from platinum and iridium, from silver and 
gold, 332, 333. Gravimetric methods, 333, 334. 

Ruthenium. — Detection — potassium hydroxide, hydrogen sulphide, ammonium 
sulphide, metallic zinc, 334. Estimation — Preparation and solution of the sample. 
Separations — ruthenium from platinum, from iridium, from rhodium, 334, 335. Gravi- 
metric methods, 335. 

Rhodium. — Detection — tests with hydrogen sulphide, potassium hydroxide, am- 
monium hydroxide, potassium nitrite, reducing agents, 336. Estimation — Prepara- 
tion and solution of the sample. Separations — rhodium from platinum, from iridium, 
from ruthenium, 336, 337. Gravimetric methods, 337. 

Osmium. — Detection — characteristics, tests with hydrogen sulphide, potassium 
hydroxide, ammonium hydroxide, reducing agents, 337, 338. Estimation — Prepara- 
tion and solution of the sample. Gravimetric methods, 338. 

Analysis of Platinum Ores, 339. Assay Methods for Platinum Ores, 340. 

POTASSIUM, SODIUM AND OTHER ALKALIES 

W. B. IIicKR, Ph.D. 
Assistant Oieinisty U. S. Geological Survey 

Detection — Sodium, 341; lithium, 342; rubidium and csesium, 342; potassium, 341. 
Estimation, 343. Solution of the sample — procedure for rocks and other insoluble 



CONTENTS xix 

mineral products, procedure for soils, fertilizers, organic compounds, ashes of plants, 
saline residues, soluble salts and brines, 343, 344. Separations — ^alkali metals from other 
constituents — ^hydrogen sulphide and ammonium sulphide groups of metals; separa- 
tion from silica; from iron, alumina, chromium, titanium, uranium, phosphoric acid; 
separation from sulphates; from barium, calcium, strontium; separation from iron, 
alumina, chromium, barium, calcium, strontium, phosphates, sulphates, etc., in one 
operation; separation from boric acid; separation from magnesium — mercuric oxide 
method, barium hydroxide method, ammonium phosphate method; separation of the 
alkali metals from one another — separation of sodium from potassium; lithium from 
sodium and potassium; lithium and sodium from potassium, rubidium and caesium, 
344-347. Methods for the determination of sodium — as sodium chloride, as sodium 
sulphate, difference method, 348, 349. Methods for the determination of potassium — 
as chloroplatinate, modified chloroplatinate method, Lindo-Gladding method, per- 
chlorate method, other methods, 349-352. Determination of sodium and potassium 
by indirect method. Determination of magnesium, sodium and potassium in presence 
of one another, 352. Methods for determining lithium — determination as lithium 
chloride, as lithium sulphate, Gooch method, Rammelsberg method, spectroscopic 
method, 353, 354. Determination of sodium, potassium, and lithium in the presence 
of one another. Determination of the alkalies in silicates by J. Lawrence Smith 
method, 355. Hydrofluoric method, 356. Determination of the alkalies in alunite, 
356. Volumetric methods, 357. 

SELENIUM AND TELLURIUM 

Wilfred W. Scott 

Detection — general procedure; detection of selenium — tests with sulphuric and 
hydrochloric acids, barium chloride, hydrogen sulphide, detection of tellurium — fusion 
test, tests with hydrogen sulphide, potassium iodide and reducing agents, 358, 359. 
Estimation — occurrence of selenium and tellurium, 359. Preparation and solution of 
the sample, selenium and tellurium solubilities, decomposition of ores, 360. Separa- 
tions — selenium and tellurium from the iron and zinc group metals, from alkalies and 
alkaline earths, from cadmium, copper and bismuth, separation from silver and gold, 
separation of selenium from tellurium by direct precipitation and by distillation 
methods, 260-262. Estimation of the two elements by the distillation separation, 
362, 363. Gravimetric methods — determination of selenium by precipitation with 
sulphmr dioxide, potassium iodide reduction method, precipitation of tellurium by 
sulphur dioxide, determination as tellurium dioxide, 264, 265. Volumetric methods — 
iodometric method for selenic and telluric acids, 265. 

SILICON 

Wilfred W. Scott 

Detection, 367. Estimation — occurrence — solubilities, 367, 368. Preparation and 
solution of the sample — general considerations, preparation of the substance for decom- 
position, general procedure for decomposing the material, silicates not decomposed by 
acids — carbonate fusion, fluorides; special procedures for decomposing the sample — 
ferrosilicons, steels containing tungsten, chromium, vanadium and molybdenum; 
silicon carbide, carborundum; sulphides, iron pyrites, slags and roasted ores, 368-371. 
Procedure for determination of silicon and silica, 372. Analysis of siHcate of soda, 
373. Analysis of sand, conunercial valuation, 374. 



CONTENTS 



SILVER 

Wallace G. Derby 
Research Chemist^ Nichols Copper Company 

Detection — wet method, silver chloride, characteristics, sundry tests, 375, 376. 
Estimation — preliminary considerations, solubility, furnace methods, 376. Gravi- 
metric methods — determination as silver chloride, as cyanide, electrolytic method, 
376, 377. Volumetric methods — thiocyanate procedure of Volhard, Gay-Lussac 
method, combination methods, Denigd's cyanide method, miscellaneous volumetric 
methods, nephelometric method, 378-384. Preparation of pure silver, 384. 

STRONTIUM 

Wilfred W. Scott 

Detection — sodium sulphate test, flame test, spectra, 387. Estimation — occurrence 
and uses, 387. Preparation and solution of the sample — solubilities, 388. Separations 
— strontium from magnesium, alkalies, calcium and barium, 388. Gravimetric methods 
— determination as sulphate, carbonate, oxide, 389. Volumetric methods — alkali- 
metric, indirect method — chloride titration with sUver nitrate, 389, 390. 

SULPHUR 

Wilfred W. Scott 

Detection — element: sulphides, sulphates, sulphites; thiosulphates, 391. Esti- 
mation — occurrence — element, sulphur dioxide, hydrogen sulphide, sulphide ores, 
sulphate ores. Preliminary considerations, 392. Preparation and solution of the 
sample — solubilities — element, sulphide, sulphate, thiosulphate, sulphite; decompo- 
sition of sulphur ores; sulphur in coal by Eschka's method; sulphur in rocks, silicates 
and insoluble sulphates, barium and lead sulphates, 292-294. Separations — sub- 
stances containing iron, separation of sulphur from metals forming an insoluble 
sulphate, nitrates and chlorates, silica, ammonium and alkali salts, 394, 395. Gravi- 
metric determination of sulphur — precipitation as barium sulphate, general con- 
siderations, precipitation from hot solutions, precipitation from cold solutions — 
large volume, standard method of the General Chemical Company, 395-398. 
Evolution method for sulphur in steel, ores, cinders, sulphides and metallurgical 
products, 398. Combustion method for evaluation of sulphide ores, 402. 
Volumetric methods for dcterminmg soluble sulphates — determination of sulphur 
by Wildenstein's method and by Hinman's method. Benzidine hydrochloride 
method, 403-405. Determination of persulphates — ferrous sulphate and oxalic acid 
methods, 406. Determination of sulphur in combination as sulphides, sulphites, 
bisulphites, metabisulphites, thiosulphates, sulphates — available hydrogen sulphide 
in materials high in sulphide sulphur — iron sulphide, sodium sulphide, etc.; hydrogen 
sulphide and soluble sulphides; sulphide and sulphydrate in presence of each other; 
thiosulphate in presence of sulphide and sulphydrate; sulphates and sulphides in pres- 
ence of one another; sulphur in thiocyanic acid and ita salts; sulphurous acid free, 
or combined in sulphites, acid sulphites, metabisulphites and thiosulphates — gravi- 
metric method by oxidation to sulphate and precipitation as barium sulphate, volu- 
metric methods — iodine titrationi acidimetric and alkalimetric methods; determination 



CONTENTS 

of sulphites, inetabisulphites, thiosulphates, sulphates, chlorides, and carbonates in 
presence of one another, 407-413. Determination of free sulphur in a mixture, 414. 
Evaluation of spent oxide for total, residual and available sulphur, 414. Analysis of 
brimstone, moisture, available sulphur, ash, arsenic and chlorine, 415. 



THORIUM 

R. Stuart Owens 
Research Chemist^ New York City 

Detection, 316. Estimation — occurrence, 416. Preparation and solution of the 
sample — silicates; phosphates (monazite). 1. By fusion with i)otassium acid sulphate; 
2. Sulphuric acid extraction; oxides, 416, 417. Separations from other elements, 417. 
Gravimetric method for determining thorium, 417. Minute amounts of thorium by 
Jolly's method, 418. 

TIN 

H. A. Baker 
Chief Chemisit A merican Can Company 

B. S. Clark 

First Assintant Chemisty American Can Company 

Detection of tin, 419. Estimation — preparation of the sample, opening up tin 
ores — the cyanide process, sodium carlwnate method, other methods — fusion with 
sodium hydrate, reduction l)y means of hydrogen, fusion with sodium peroxide, 419- 

421. Separations — general procedure, separation of tin from lead, copper, antimony, 
phosphorus, iron and alumina, tungstic acid, 421, 422. Gravimetric methods — de- 
termination of tin or the oxides of tin by hydrolysis, determination of tin as sulphide, 

422, 423. Bichloride of tin — stannic acid method — hot water precipitation; Acker 
process method; determination as sulphide, 424-426. Volumetric determination of 
tin — Lenssen's iodine method as modified by Baker, standard method of the American 
Can Company, 426. Electrolytic determination of tin, 430. Estimation of tin in 
canned food products, 430. 

TITANIUM 

Wilfred W. Scott 
Research Chemist^ General Chemical Company 

L. E. Barton 
Chief Chemist J Titanium Alloy Manufacturing Company 

Detection — tests with hydrogen peroxid:, morphine, zinc, sulphur dioxide and 
by fusion with microcosmic salt, 432. Estimation — occurrence and application, 433. 
Preparation and solution of the sample — element, oxides, salts of titanium — general 
considerations — solution of steel, alloys, ores, titaniferous slags, 433, 434. Separations 
— titanium from the alkaline earths, etc.; separation from copper, zinc, aluminum, 
iron, manganese, nickel, cobalt, 434, 435. Gravimetric methods — modified procedure 
of Gooch, 435. Determination of titanium in ferro-carbon titanium, 436. Volumetric 
methods — reduction of titanic solution, addition of ferric salt and titration of reduced 
iron with permanganate; reduction of titanic salt and titration with ferric salt, 437-439. 



xxu CONTENTS 

Colorimetric detennination of titanium with hydrogen peroxide; coloiimetric deter- 
mination in steel treated in ferro-carbon titanium, (a) determination of titanium 
insoluble in hydrochloric acid, (6) determination of titanium soluble in hydrochloric 
acid, (c) total titanium; determination in presence of interfering elements; colori- 
metric determination of titanium with th3rmol solution. Analysis of titaniferous 
ores— determination of titanium, silica, alumina, phosphorus. Standard methods of 
the Titanium Alloy Manufacturing Company, 439-447. 

TUNGSTEN, TANTALUM AND COLUMBroM 

Wilfred W. Scott 

Tungsten. — Detection — minerals, iron, steel and allo3rs, 448. Estimation — 
occurrence and uses, 449. Solution of the sample — solubilities, decomposition of ores, 
acids, minerals, steel and alloys, steel containing a high percentage of tungsten, ferro- 
tungsten alloys, tungsten bronzes, 449-451. Separations — tungsten from sihca, 
separation of tungsten from tin, antimony by Talbot's process, separation from arsenic 
and phosphorus, separation from molybdenum by Hommel's process, volatilization of 
molybdenum with dry hydrochloric acid gas by Pochard's process; separation from 
vanadium, titanium, iron, 451-453. Gravimetric procedures — precipitation of tung- 
stic acid, precipitation as mercurous tungstate by BerzeUus' process, 453, 454. Volu- 
metric method, 454. 

Tantalum and Columbium. — Detection, 455. Estimation — occurrence, applica- 
tion, 455. Solution of the sample — general procedure, tantah'ferous minerab, 456. 
Separations — isolation of columbium and tantalum oxides; removal of tin, antimony, 
tungsten and silica, 456, 457. Determination of columbium and tantalum, 457. 

URANIUM 

Wilfred W. Scott 

Detection — general procedure — uranous salts, uranyl salts, 458. Estimation — 
occurrence, industrial appUcation, 458. Preparation and solution of the sample — 
solubilities — element, oxide, salts; solution of ores, 459. Separations — uranium 
from copper, lead, bismuth, arsenic, antimony, and other members of the hydrogen 
sulphide group; separation of uranium from iron and from elements having water 
insoluble carbonates; separation from vanadium, 459, 460. Gravimetric determina- 
tion of uranium as the oxide, UaOs, 461. Volumetric determination of uranium by 
reduction and subsequent oxidation, 461. 

VANADIUM 

Wilfred W. Scott 

Detection — tests with sulphide, reducing agents, hydrogen peroxide, ammonium 
chloride, distinction from chromium, detection in steel, 463, 464. Estimation — 
occurrence and industrial application, 464. Preparation and solution of the sample — 
solubility of fhe element, its oxides and salts; general procedure for decomposition of 
ores of vanadium, ores high in silica, products low in silica, iron, steel and alloys, 465, 
466. Separations — general procedure, removal of arsenic, molybdenum, phosphoric 
acid, separation of vanadium from chromium, 466, 467. Gravimetric methods— deter- 
mination of vanadium by precipitation with mercurous nitrate; by precipitation with 



CONTENTS xxiii 

lead acetate, 467, 468. Volumetric methods — reduction to vanadyl condition and 
oxidation with potassium permanganate, reduction with zinc followed by perman- 
ganate titration; determination of vanadium in steel; determination of molybdenum 
and vanadium in presence of one another; determination of vanadium, arsenic or 
antimony in presence of one another by Edgar's method; determination of vanadium 
and iron in presence of one another; iodometric method for estimation of chromic 
and vanadic acids in presence of one another; determination of vanadium in ferro- 
vanadium, methods of the Vanadium Company of America — general procedure, vana- 
dium in ores; in steel; in steel containing chromium; in cupro- vanadium; in brasses 
and bronzes, 46^-476. 

ZINC 

F. G. Breyer, M.A. 
Chief of the Testing Department ^ New Jersey Zinc Company (of Pa.) 

Detection of zinc, 477. Estimation, 477. Preparation of the sample — moisture 
determination in the pulp, 478. Separations — from silica, cadmium, arsenic, 
antimony, bismuth, copper, iron, alumina, manganese, nickel and cobalt, 478, 
479. Methods of analysis — Gravimetric methods — weighing as the oxide, electrolytic 
procedure, 479. Volumetric methods — ferrocyanide titration of the acid solution, 
separating iron, aluminum and manganese with ammonia and bromine; titration 
of the alkaline solution — procedure for comm:)n ores; procedure for copper-bearing ores; 
procedure for material containing cadmium; for material containing carbonaceous 
matter; procedure for material containing mctallics; general notes, 480-483. Standard 
method of the New Jersey Zinc Company — titration in acid solution — separating of 
zinc as sulphide; standardization of the ferrocyanide solution; procedure with material 
containing insoluble zinc; discussion on separating zinc as zinc sulphide and titrating 
in acid solution, 483-487. Determination of small amounts of zinc, 487. Special 
methods — determination of metallic zinc in zinc dust, 487. Determination of impurities 
in spelter-lead by electroljrtic ::nd ** lead acid " methods; iron by colorimetric and hydro- 
gen sulphide methods; cadmium by sulphide and electrolytic method, 489-492. 
Determination of impurities in zinc oxide (reference) ; general references, 492, 493. 

ZIRCONIUM 

R. Stuart Owens 

Research Chemist, New york City 

Detection, 494. Estimation, 494. Preparation and solution of the sample — 
materials containing a large amount of silica, general method for minerals, oxides, etc., 
other methods, 494, 495. Separations — from iron, titanium, thorium, carium, the 
iron group, 495. Gravimetric methods for the determination of zirconium — salts of 
zirconium, minerals and silicates; determination as phosphate; determination as 
zirconium oxide; determination as oxide in presence of iron, 496. 



xxiv CONTENTS 



PART II 

SPECIAL SUBJECTS 

Acms 

Wilfred W. Scott 

Indicators — classification and special uses of, 499-501. Ultimau; standards — 
preparation of pure sodium carbonate, 501. Preparation of standard acids — sulphuric 
acid, hydrochloric acid, benzoic acid, standard caustic solution, 502-504. Standard 
burettes, 505. Methods of weighing acids — dilute acids non-volatile under ordinary 
conditions; weighing of strong acids, fuming or volatile under ordinary conditions; 
Lunge-Ray pipette, Dely weighing tube, snake weighing tube, Dlay-Burkhard grad- 
u::ted weighing burette, 506, 508. Titration of acids and alkalies, 508. Analysis 
of muriatic acid — total acidity and hydrochloric acid. Impurities in commercial 
hydrochloric acid — free chlorine, nitric acid or nitrates; sulphuric acids and sulphates, 
arsenic, barium chloride, silica and total solids, 509, 510. Analysis of hydrofluoric 
acid — total acidity, hydrofluosilicic acid, sulphuric acid, sulphurous acid, calculation 
of results, 510-512. Complete analysis of nitric acid — total acidity, sulphuric acid 
hydrochloric acid, lower oxides, nitric acid, iodine, free chlorine, total non-volatile 
solids, 512-514. Ferrous sulphate method for the direct determination of nitric acid — 
standardization of the reagents; general procedure for nitric in sulphuric acid; evalua- 
tion of nitric acid or nitrates; determination of nitric acid in oleum or mixed acids; 
determination of nitric acid in arsenic and in phosphoric acids, 515-520. Determina- 
tion of nitrous acid or nitrite by the permanganate method, 520. The analysis of 
oleum or fuming sulphuric acid and of mixed acid — total acids, lower oxides, sulphuric 
acid and free sulphuric anhydride, nitric acid, calculating of results, 522-526. Analysis 
of acetic acid — impurities in acetic acid, formic acid, furfurol, acetone, sulphuric acid, 
sulphurous acid, hydrochloric acid, metals in acetic acid, 527-529. Acetates, 529. 
Citric acid, 530. Volumetric estimation of free acid in presence of iron salts, 530. 
Estimation of carbonates and hydrates of potassium and sodium when together in 
solution, 531. 

WATER ANALYSIS 

D. K. French 

Director of the Laboratory^ Dearborn Chemical Company 

General considerations, 533. Sanitary analysis — organic nitrogen; chloride; 
oxygen consumed; physical test — turbidity, color, odor — hot or cold; taste, 534-535. 
Chemical tests — free ammonia, albuminoid anmaonia, organic nitrogen; nitrogen as 
nitrite; nitrogen as nitrate by phenolsulphonic acid method and by aluminum reduc- 
tion; ox}'gen consumed; chlorine as chlorides; total soHd residue, 536-542. Inter- 
pretation of results, 543. Mineral analysis — general considerations, outline of pro- 
cedure, silica, manganese and phosphoric acid; iron and alumina gravimetric; iron 
colorimetric, ferrous iron colorimctric; phosphates, calcium, magnesium, manganese — 
Knorre's persulphate method, sodium bismuthate method for manganese; sulphates — 
benzidine method; sodium and potassium; alkalinity, acidity, free carbonic acid; 
chlorine; nitrates; ammonia and its compounds; total mineral residue; hydrogen 
sulphide; oil; dissolved oxygen by Winkler's method, 545-557. Methods for deter- 



CONTENTS XXV 

mining small amounts of lead, zinc, copper and tin, 557. Hardness, preparation of 
solutions, magnesium chloride, calcium sulphate; lime and soda value, 558-561. 
Methods of reporting and interpretation, 562. Water softening, foaming and prim- 
ing, corrosion, scale, irrigating waters, hypothetical combinations, 563-565. Field 
assay of water, 565. 

FIXED OILS, FATS AND WAXES 

Augustus H. Gill, Ph.D. 
Profeswr of Technical Analysis, Massachusetts Institute of Technology 

Introductory. Examination of an unknown oil, 566. Petroleum products — 
Burning oils; flash test, determination by the New York State Board of Health tester; 
fire test; specific gravity, (a) by the hydrometer, (6) by the Westphal balance; dis- 
tillation test, Engler's method; determination of sulphur; detection of acidity; sul- 
phuric acid test; mineral salts; determination of water; color, 567-572. Lubricating 
oils; viscosity, Engler apparatus, Saybolt Universal viscosimeter, absolute viscosity; 
specific gravity; evaporation test; cold test; flash point; fire test; detection of 
soap; caoutchouc; tests for fatty oils; gumming test; carbon residue test, 
Gray's method; gasoline test; microscopical test; friction'tests, 572-580. Animal 
and vegetable oils — specific gravity; refractive index; Valenta test, elaidin test; 
Maumeni^ test; iodine number; Hanus's method, Hiibrs method; oxidized oils, iodine 
number; bromine number; saponification value; detection of unsaponifiable oils, 
from saponification number, by gravimetric methods; identification of the unsaponi- 
fiable matter; test for animal or vegetable oils; tests for antifluorescents; acetyl 
value, 580-591. Special tests for certain oils — Bechi's test for cotton-seed oil; Hal- 
pen's test for cotton-seed oil; hexabromide test for linseed oil; Renard's test for pea- 
nut oil; Bach*s test for rapeseed oil; Licbermann-Storch test for rosin oil; Baudouin's 
or Camoin's test for sesam6 oil; free acid test; spontaneous combustion test, Mackey*s 
apparatus; drying test upon glass; titer test, 591-599. Edible fats — butter, examina- 
tion of the fat; preservatives, color; lard, water, 599-600. Hardened oils, 600. Waxes, 
601. Miscellaneous oils and lubricants — general description, drying, semidrying and 
non-drying oils, 601-602. Tables — properties of some mineral oils; characteristics 
of the fatty acids from some oils; characteristics of some oils; characteristics of some 
waxes, 601-607. Reagents, 607, 608. 

ANALYSIS OF PAINTS 

Henry A. Gardner 
Asaistani DiredoTf The Institute of Industrial Research, Washington, D, C. 

John A. Schaeffer, Ph.D. 
Chief Chemist, Ecujle-Picher Lead Company, Joplin, Mo, 

Introductory, 609. Analysis of paint vehicles — composition of liquid part; per- 
centage of liquid by ignition method; percentage of liquid by extraction methods; 
separation of vehicle components, water, direct distillation of volatiles; detection of 
resinates; detection of various oils, 610-612. Analysis of paint oils — iodine number; 
analysis of Chinese wood oil (tung oil), specific gravity, add number, saponification 
number, unsaponifiable matter, refractive index, iodine number (HUbl), heating 
test (Browne's method), iodine jelly test; standards of Chinese wood oil, A. S. T. M.; 



xxvi CONTENTS 

constants of various oils, comparative tables; examination of turpentine — color, specific 
gravity, refractive index, distillation, polymerization; standards for turpentine, 
A. S. T. M., 612-618. Analysis of varnish — flash point, acid number, ash, solvent, 
fixed oils and resins, separations'of polymerized oils and resins, 618-620. Other mate- 
riab, 620. The analysis of paint pigments; classification of pigments, 621. Analysis 
of white pigments — sublimed white lead, volumetric determination of lead, volumetric 
determination of zinc, total sulphate; corroded white lead, total lead (gravimetric), 
total lead (volumetric), carbon dioxide, acetic acid, metallic lead; zinc lead and leaded 
zinc — moisture, lead, zinc, total soluble sulphates (in absence of BaS04), total 
soluble sulphates (in presence of BaS04), soluble zinc sulphate, sulphur dioxide, cal- 
culations; zinc oxide — moisture, carbon dioxide, insoluble matter, sulphuric anhydride, 
total S as SOs; lead oxide, gravimetric method, electrolytic method, chlorine, ferric 
oxide, manganese oxide, arsenous oxide, SOi equivalent, zinc oxide; lithopone — 
moisture, barium sulphate, total zinc, zinc sulphide, soluble salts; silex — moisture, loss 
on ignition, insoluble matter, carbon dioxide; whiting — Paris white — ^moisture, loss 
on ignition, calcium, magnesium, carbon dioxide, sulphates; barytes and blanc 
fixe — moisture, loss on ignition, bariimi culphate, soluble sulphates, carbon dioxide; 
analysis of a composite white paint — insoluble residue, total lead, alumina and iron 
oxide, zinc, calcium and magnesium, sulphate, sulphide, carbon dioxide, calculations, 
622-634. Red and brown pigments — red lead and orange mineral — moisture, organic 
color, total lead and insoluble residue, lead peroxide (PbOj) and true red lead (PbaO^), 
calculation; vermilion — characteristics of; iron oxides, 634, 637. Blue pigments — 
ultramarine blue — moisture, silica, aluminum oxide, sodium oxide, total sulphur, sul- 
phur present as sulphate; Prussian blue — moisture, nitrogen, iron and aluminum oxides, 
sulphuric acid, conmiercial anal3rsis; sublimed blue lead — ^total lead, total sulphur, 
lead sulphate, lead sulphite, lead sulphide, lead carbonate, lead oxide, zinc oxide, carbon 
and volatile matter, 637-639. Yellow and orange pigments — chrome yellow, mois- 
ture, insoluble residue, lead, chromium, zinc, calcium and magnesium, sulphuric acid, 
calculations, 639. Green pigments, chrome green — moisture, insoluble residue, lead, 
iron, alumina and chromium, calcium and magnesium, sulphuric acid, nitrogen, cal- 
culation, 639, 640. Black pigments — moisture, oil, carbon, ash, analysis of ash, 640- 
641. Ck>mplex compounds — hydroferrocyanic and hydroferricyanic acids, 641. 

CEMENTS 

Richard K. Meade, M.S. 

Chemical, Mechanical and Industrial Engineer, Baltimore, Md, 

Analysis and testing of cements — introductory, 642. Physical testing — ^fineness, 
specific gravity, normal consistency, table — percentage of water for standard sand 
mortar, setting time, soundness or consistency of volume, tensile strength, notes, 
apparatus for testing of cement, 642-649. Standard method for chemical analysis of 
Portland cement — solution, silica, alumina and iron, lime, magnesia, alkalies, anhydrous 
sulphuric acid, total sulphur, loss on ignition, insoluble residue, 650, 653. Rapid 
method for chemical anfd3rsis of Portland cement, 653. Rapid method for checking 
the percentage of calcium carbonate in cement mixture, standard alkali, standard acid, 
standard sample, standardizing the acid, determination, 656-658. Analysis of lime- 
tone, cement rock, lime, Roaendale cement, etc., 658. 



CONTENTS xxvii 

ANALYSIS OF ALLOTS 

John C. Olsen, Ph.D. 

Professor in Charge of the Department of Chemistry y Cooper Unions New York City 

Introduction — difficulty of complete separation of elements, limit of accuracy in 
analysis, 659. Analjrsis of tjrpe metal — solution of the alloy, lead, copper and iron, 
separation of antimony and tin, determination of antimony, tin and arsenic, 650-661. 
Analysis of soft solder — solution of the alloy, determination of tin, lead, arsenic, 
antimony, iron, zinc, 661-663. Analysis of Rose's metal — decomposition of, material, 
determination of lead, bismuth, copper, 663. Analjrsis of Wood's metal — decomposi- 
tion of the material, determination of lead, bismuth, cadmium, arsenic, tin, separation 
of copper and cadmium, determination of copper, separation and estimation of iron 
and zinc, 664, 665. Analysb of Britannia metal — decomposition of the alloy by means 
of chlorine, determination of lead, copper, iron, bbmuth, separation of tin from arsenic 
and antimony, determination of tin, arsenic and antimony, 666, 667. Analysis of 
brass or bronze — solution of the alloy, determination of tin, arsenic, antimony, lead, 
copper, iron, zinc, 667-669. Analysis of German silver — decomposition of the mate- 
rial, determination of zinc, iron and nickel, 668. Analysis of manganese and phos- 
phorous bronze — solution of the alloy, determination of lead, copper, zinc, iron, man- 
ganese, phosphorus, 670, 671. 

METHODS FOR ANALYSIS OF COAL 

Frank E. Hale, Ph.D. 

Director of Laboratories, Department of Water Supply, Gas and Electricity, New York City, 
Sampling; preparation of the sample for analysis, 672-674. Methods of analysis 
— ^moisture, ash, volatile combustible matter, volatile sulphur, turbidimetric sulphur 
table, fixed carbon, calorific value, calculation of B.t.u., standardization of the calo- 
rimeter, 674-683. Determination of fusibility of coal ash, 684. References, 685. 

GAS ANALYSIS 

Augustus H. Gill, Ph.D. 

Professor of Technical Analysis, Massachusetts Institute of Technology 

Sampling — tubes, pumps, containers for samples, 687-689. Measurement of 
gas in large quantities — ^wet and dry meters, Pitot tube or Davis anemometer, rotam- 
eter or Thorpe gauge, capometer, Thomas electric meter, orifice meter, anemometer, 
689-692. Measurement of gas in small quantities — gas burettes, Hempel Orsat 
and Elliot burette, separatory funnel and graduate, 692. Absorption apparatus, 
tubes and pipettes, 693. Examination of the gases — detection and determination 
of the various gases, tables, 694-697. Analysis of gaseous mixtures — analysis by means 
of the Orsat apparatus, determination of carbon dioxide, oxygen, carbon monoxide, 
hydrocarbons, notes on manipulation; Elliot apparatus, determination of carbon 
dioxide, oxygen, carbon monoxide, notes; Hempel apparatus, determination of oxygen 
in air — (1) by phosphorus, (2) by p3ax)gallate of potassium, (3) by explosion with 
hydrogen; analysis of illuminating gas — carbon dioxide, illuminants, oxgyen, carbonic 
oxide, methane and hydrogen, (a) Hinman's method, (&) Hempel's method; nitrogen, 
notes, 697-707. Applications of gas analysis and interpretation of results — I. Chimney 



xxviii CONTENTS 

and flue gases — carbonic acid indicators, determination of temperature, composition 
of the coal, tables, smoke, 708-711. II. Producer and fuel gases, blast-furnace gas — 
analysis, dust in, 711-712. III. Illuminating gas, candle-power, calorific power, sulphur, 
H2S, ammonia, naphthalene, carbon dioxide, specific gravity, tar, 712-720. IV. Sul- 
phuric acid gases — (a) burner gases — sulphur dioxide in inlet and exit gases, Reich 
method for SOs, absorption of SO2 in chromic acid solution; (&) nitrogen oxides, 720- 
726. V. Mine gases, 726. VI. Electrolytic gases, 727. VII. Acetylene, 727. Atmos- 
spheric air moisture, carbon dioxide, ozone, carbon monoxide, bacteria, 728-730. 
Determination of moisture in gases, 731. Determination of nitrogen by the nitrometer, 
732. Reagents and tables, 734-738. 

ASSAYING 

Wallace G. Derby, M.S. 

Research Chemisij Nichols Copper Company 

Sample, the unit of weight, general survey of the subject of sampling, 739. 
Furnace operations; consequent to the furnace operations; preliminary to the furnace 
operations; silver and gold retained in the slag; silver and gold retained in the cupel; 
corrected assay; determination of gold; determination of silver; influence of quantity 
of sample, 741. Roasting, incineration, 743. Crucible method of fusion, 744. Scori- 
fication method of fusion, 755. Cupellation, 759. Parting, 766. Combination meth- 
ods, 769. Determination of gold in cyanida solution, 773. 

PART III 

TABLES AND USEFUL DATA 

I. International atomic weighta, 779. II. Melting-points of chemical elements, 
780. III. Temperature standards, 780. IV. Electromotive arrangement of the ele- 
ments, 781. Specific Gravity Tables of the Acids and Alkalies, 782-799. 
V. Hydrochloric acid — Ferguson, 782. VI. Hydrochloric acid — Lunge and March- 
Icwski, 784. Constant boiling-points, 784. VII. Nitric acid, Ferguson, 785. VIII. 
Nitric acid. Lunge and Rey, 787. IX. Phosphoric acid — Hager, 789. X. Sul- 
phuric acid — Ferguson and Talbot, 790. XI. Sulphuric acid — Bishop, 794. XII. 
Acetic acid — Oudemans, 795. XIII. Melting-points of acetic acid — Rudorff, 795. 
XIV. Aqua ammonia — Ferguson, 796. XV. Sodium hydroxide — Lunge, 798. XVI. 
Vapor tension of water in milligrams of mercury, from —2® to +36** C. — Regnault, 
Broch and Weilje, 800. XVII. Useful data of the more important inorganic com- 
pounds — Meiklejohn, 801. XVIII. Conversion factors — Scott and Clark, 804. XIX. 
Comparison of centigrade and Fahrenheit scale, 818. XX. Relation of Baum4 
degrees to specific gravity and the weight of one U. S. gallon at 60® F., 819. XXI. 
Comparison of customary units of weight and measure with the metric system, 
820. XXII. Table of constants for certain gases and vajwrs, 822. XXIII. Solu- 
bility table, 824. Tables of Qualitative Tests, 825-855: XXIV. Blowpipe and 
flame tests of solids, 827-829. XXV. Separation of the bases — Analysis of the solu- 
tion, 830, 831. XXVI. Tests for acids, 832. XXVII. Tables of reactions of the 
bases, 834-847. XXVIII. Tables of reactions of the acids, 848-855. General refer- 
ences, 856. Index, 861-898. 



LIST OF ILLUSTRATIONS 

PAGB 

1. Apparatus for Distillation of Arsenous Acid 34 

2. Urbasch's Hydrogen Sulphide Generator 37 

3. Scott's Hydrogen Sulphide Generator 38 

4. Banks' Hydrogen Sulphide Generator 39 

5. Purification of Hydrochloric Acid 41 

6. Gutzeit Apparatus for Arsenic Determination 46 

7. Purification of Carbon Disulphide 67 

8. Distillation of Methyl Borate 74 

9. Test for Carbonate 93 

10. Chilled Steel Mortar 94 

11. Geissler Bulb 96 

12. Liebig Bulb 96 

13. Gerhardt Bulb 96 

14. Vanier Bulb 96 

15. Fleming's Apparatus for Determination of Carbon by Combustion 97 

16. Fleming's Absorption Apparatus 98 

17. Boat and Holder for Carbon 100 

18. Shimer Combustion Apparatus 100 

19. Diagrammatic Sketch of Combustion Tul)e 101 

20. Apparatus for Determining Carbon Dioxide 104 

21. Shroetter's Alkalimeter 106 

22. Mohr's Alkalimeter 106 

23. 24. Hot Water Racks for Test Tubes; Color Carbon Determination 108 

25. Carbon Tubes 109 

26. Color Comparitor or Camera 109 

27. Terminal Case Showing Battery of Electrodes 155 

28. Solenoid for Rotation of Electrolyte 157 

29. Riffle Sampler 159 

30. Constant Temperature Bath and Dividing Pipette 160 

31. Hydrometer Jar for Electrolysis of Copper 160 

32. Special Beaker for Electrolysis of Copper 173 

32a. Combustion Furnace, Hinged Type 174 

33. Etching Test for Fluorine 176 

34. Hanging Drop Test for Fluorine 177 

35. Black Filter Paper Test for Fluorine 177 

36. Adolph's Apparatus for Determining Fluorine 182 

37. Steiger-Merwin Fluorine-Titanium Chart 185 

38. Merwin's Chart on Ratio of Depth of Color 185 

39. Apparatus for Determining Iodine in Iodide 206 

40. 49. Jones' Reductor 220, 281 

41. Apparatus for Stannous Chloride Titration of Iron 221 

42. Dividing Pipette 223 

xxix 



XXX LIST OF ILLUSTRATIONS 

PIG. PAGS 

43. Hurley's Colorimeter 245 

44. Cooper Hewitt's Mercury Light 247 

45. Bell Jar Vacuum Filtering Apparatus 265 

46. Automatic Measuring Pipette 265 

47. Sulphur Extraction Apparatus 272 

48. Holloway-Eschka Apparatus for Determining Mercury 273 

60. Apparatus for Determining Nitrogen, Kjeldahl Method 294 

61. Devarda's Apparatus Modified by Allen 301 

62. Weighing Bottle and Dropper 302 

63. Mitscherlich's Apparatus, Nitrogen in Soils 305 

64. Nitrometer 307 

65. Du Font's Nitrometer 307 

66o. Becker Chain Balance 323 

66. J. Lawrence Smith Apparatus 355 

67. Apparatus for Separation of Selenium and Tellurium 362 

58. Lead Cup for Silica Test ; 367 

59. Apparatus for Determining Silver, Gay-Lussac Method 381 

60. Evaporation in Sulphur Determination 397 

61. 62. Apparatus for Precipitating Sulphur 397 

63. Apparatus for Filtering Barium Sulphate 398 

64. Scott's Apparatus for Determining Sulphur in Iron and Steel 399 

65. Arrangement for Protecting the Crucible from the Flame 401 

66. Sanders' Extraction Apparatus 414 

67. Seller's Apparatus for Determining Tin 428 

68. Seller's Apparatus, Diagrammatic Sketch 429 

69. 70. Colorimeter 440, 442 

71. Voit Flask and Distillation Apparatus 473 

72, 73, 74. Apparatus for Determining Zinc 488 

75. Arrangement for Heating Sodium Bicarbonate 501 

76, 77, 78. Charts Showing Specific Gravity and Boiling-points of Sulphuric 

Acid of Varying Concentration 502 

79. Standard Burette 505 

80. Lunge-Ray Pipette 506 

81. Dely Weighing Tube in Operation 507 

82. Snake Tube 507 

83. Blay-Burkhard Graduated Weighing Burette 508 

84. Method for Rapid Evaporation of Liquids 522 

85. Apparatus for Determining Ammonia in Water 536 

86. New York Tester 567 

87. Westphal Balance 569 

88. Englcr Viscosimeter 573 

89. Saybolt Viscosimeter 573 

90. Cleveland Cup 577 

91. Gray's Distillation Flask 579 

92. Refractometer 581 

92a. Mackay's Apparatus 596 

93. Le Chatelier's Specific Gravity Apparatus 643 

94. Vicat Needle 644 

95. Gilmorc Needles 645 



LIST OF ILLUSTRATIONS xad 

Fia. PAOB 

96. Pat for Determining Setting Time and Soundness in Cement 646 

97. Appearance in Pats Made from Soimd and Unsound Cement after Steaming 646 

98. Details for Briquette 647 

99. Details for Gang Mold 647 

00. Fairbanks^ Cement Testing Machine 648 

01. Riehl6 Automatic Cement Testing Machine 649 

02. Apparatus for Determining Calcium Carbonate with Acid and Alkali 657 

03. Quartering Coal, Ball Mill for Pulverizing, and Suction Ventilator 673 

04. V. C. M. Apparatus 675 

05. Atwater Bomb and Calorimeter 679 

06. Oxygen Cylinders for Calorimeter 680 

07. Haskins' Electric Furnace, Optical Pyrometer Outfit 684 

08. Sampling Tube for Gas 687 

09. 110, 111. Pumps 688 

12. Container for Gas Sample 689 

13. Pitot Tube 690 

14. Rotameter 691 

15. Capometer 691 

16. Friedrichs' Spiral Gas Washing Bottle 693 

17. Varentrapp and Will Bulbs 693 

18. Wolf Absorption Bulb 693 

19. Winkler's Spiral 693 

20. Orsat Apparatus 697 

21. Elliott's Apparatus 700 

22. Hempel's Apparatus 702 

23. Hempel's Combustion Apparatus 703 

24. 125, 126. Junker's Calorimeter. 713, 714 

27. Apparatus for Determining Sulphur in Gas 716 

28. Rudorff's Apparatus for COa in Gas 719 

29. Specific Gravity Apparatus for Gas 719 

30. Reich Apparatus for SOa in Contact Gas 721 

31. Briggs-Scott Modified Orsat Apparatus for SO2 in Contact Gas 723 

32. Hesse's Apparatus for COa in Air 728 

33. Absorption Spectrum Chart, CO2 Determination 730 

34. Phosphorus Pentoxide Bulb for Water Vapor in Gas 731 

35. Apparatus for Determining Gasoline Vapor in Gas 731 

36. Nitrometer 732 

37. Bunte's Chart 738 

38. Chart Showing Influence of Quantity of Gold or Silver in Assaying 743 

39. Assay Furnace 744 

40. Assaying Outfit 755 

41. Chart Showing Cupellation Loss 764 

42. Assay Balances 767 

Plate I. Arsenic Stains, Gutzeit Method Facing page 46 

Plate II. Emissions Spectra Facing page 53 

Plate III. Diffraction Grating Spectrum and Prismatic Spectrum. . . .Facing page 341 



PART I 

TECHNICAL METHODS FOR THE DETECTION AND 
DETERMINATION OF THE MORE IMPORTANT 

ELEMENTS 



ALUMINUM » 

Wilfred W. Scott 
AU at.wt. 27.1; 8p.gr. 2.583; m.p. 658.7*"; b.p. 2200''; oxide AlsOt. 

DETECTION 

The sample is brought into solution according to one of the procedures out- 
lined under " Preparation and Solution of the Sample." Silica is removed by 
taking the solution to dryness, boiling the residue with hydrochloric acid and 
filtering. The members of the hydrogen sulphide group are removed as usual 
with HjS, the filtrate boiled to expel the excess of HjS, iron oxidized with nitric 
acid, and aluminum, iron and chromimn precipitated as hydroxides by addition 
of ammonium hydroxide in presence of ammonium chloride. On treating the 
precipitate with sodium peroxide, aluminum and chromium hydroxides dissolve, 
whereas ferric hydroxide remains insoluble. Aluminum hydroxide is precipi- 
tated by acidifying the alkaline solution with hydrochloric or nitric acid, and 
neutralizing with ammonia; chromium remains in solution. 

The white gelatinous precipitate of ahuninum hydroxide may be confirmed 
by adding a drop of cobalt nitrate solution and burning the filter. The 
residue will be colored blue by the resulting aluminum cobalt compound. 

Sodium thiosulphate, NajSiOa, added to a neutral or slightly acid solution, 
containing almninum, precipitates aluminum hydroxide, upon boiling the solu- 
tion. Sodiimi sulphite, or ammoniimi chloride added in large excess, will also 
cause this precipitation. 

ESTIMATION 

The determination of aluminum, in terms of alumina, AI2OS, is required in 
the evaluation of aluminum ores, bauxite, AUOCOH)*; diaspore, AIO(OH); alunite, 
K:0.3Al20i.4SOj.6H20, etc. It is determined in the analysis of feldspar, hal- 
loysite, clays, granite, gneiss, porphyry, mica schist, slate, obsidian or pumice 
stone, cryolite, limestone, and in the complete analysis of a large number of 
mineral substances. The estimation of alumina is required in the analysis of 
cements, plaster, ceramic materials, aluminum salts, and is especially important 
in the control of processes in the manufacture of aluminum products. As a 
metal it is determined in conmiercial aluminum, and its alloys. 

Preparation and Solution of the Sample 

In dissolving substances containing aluminum it will be recalled that alumina, 
although ordinarily soluble in acids, is very difficult to dissolve when it is highly 
heated. It may be best dissolved, in this case, by fusion with sodium carbonate 
or with acid potassium sulphate, followed by an acid extraction. The metal is 
scarcely acted upon by nitric acid, but is readily soluble in the halogen acids and 
in hot concentrated sulphuric acids. 

General Procedure for Ores. One gram of the finely powdered ore, 
taken from a representative sample, is placed in a platinimi dish, 5 cc. of con- 

^ Also spelled Aluminium. 
3 



4 ALUMINUM 

centrated Bulpfauric acid are added, followed by about 20 cc. of strong hydro- 
fluoric acid. The mixtiu'e is evaporated over a steam bath as far as possible 
and then taken to SO3 fumes on the hot plate {Hood). Upon cooling, a little 
dilute hydrochloric acid is added and the mixture warmed. The solution is 
diluted with distilled water and filtered if any residue remains. 

The insoluble residue remaining on the filter may be brought into solution 
by fusing the ignited residue with sodium carbonate or acid potassium sulphate. 
If barium is present sodium carbonate fusion is made and the melt extracted 
with water to remove the sodium sulphate. The residual carbonates may now be 
dissolved with hydrochloric acid. 

Sulphide Ores should be oxidized with nitric acid and bromine according 
to the general procedure for decomposing pyrites in the determination of 
sulphur. 

The solution of the sample having been effected, aluminum is separated from 
elements that interfere in its estimation. Directions for the removal of these 
substances are given under " Separations." The element is now in solution in 
such form that it may be determined gravimetrically or volumetrically. 

Fusion Method. Sodium Carbonate. The air-dried material, ground to a 
fine powder, is placed in a glass-stoppered bottle. If the determination is to be 
made on the dry basis, moisture is driven out by placing the material in the hot air 
or steam oven for an hour (100 to 110° C). One gram sample, placed in a large 
platinum crucible, is mixed with 4 to 5 grams of anhydrous sodium carbonate 
and the material heated to fusion, the heating being continued until the molten 
mass appears clear. The hquid mass may be poured on a large platinum crucible 
lid, or if preferred, allowed to cool in the crucible, a platiniun prod being held in the 
fusion until it solidifies. By gently heating the crucible over a flame the fusion 
loosens from the sides and may be lifted out on the prod. In either case the 
cooled mass is dissolved by placing it, together with the crucible in which the fusion 
was made, in a casserole, and treating with hydrochloric acid, the casserole being 
covered with a clock glass during the reaction. 

SiUca is removed by evaporating the solution to dryness on the water or steam 
bath and drying in the oven at 110° C. for an hour or more. The residue is 
extracted with hot dilute hydrochloric acid and silica filtered off. 

If the solution is cloudy upon treatment of the fusion with acid, it indi- 
cates either the presence of barium sulphate or incomplete decomposition of 
the sample. In the latter case the residue is gritty and the fusion of this 
material should be repeated. 

Fusion with Potassium Bisulphate. This procedure is recommended for 
the decomposition of commercial alumina or calcined Al20a. The procedure is 
similar to the sodium carbonate fusion with the exception that less heat is required. 
A silica dish may be used, if desired, in place of platinum. 

Extraction of Ores of Aluminum for Their Commercial Valuation. The 
available alumina in bauxite, alunite, clays and aluminum-bearing materials 
may be approximately ascertained by digesting 5 grams of the pulverized sample 
with 45 cc. of 5 N. H2SO4 for three hours in a Kjcldahl flask with reflux con- 
denser, the heat being so regulated ^ that the drip from the condenser amounts 
to 5 to 8 drops per minute. The solution filtered hot, together with the wash- 
ings of the residue is diluted to 1000 cc. Aliquot portions of this solution 

^ Method for controlling temperature by observing condensation suggested by 
W. S. Allen. 



ALUMINUM 5 

are taken for determination of the desired soluble constituents, according 
to the procedures outlined under bauxite analysis in the latter portion of the 
chapter. 

Solution of Metallic Aluminum and its Alloys. The metal may be dis- 
solved in dilute hydrochloric acid, 1 : 1, or in a solution of sodium hydroxide 
or potassium hydroxide. 

Alloys of aluminum are best brought into solution with a mixture of hydro- 
chloric and nitric acids. 

SEPARATIONS 

General Considerations. In the usual course of analysis, aluminum is 
in solution as a sulphate or as a chloride, silica having been removed by dehy- 
dration, as described under "Preparation and Solution of the Sample." The 
following interfering elements may be present in the solution: iron, manganese, 
arsenic, antimony, titanium, phosphoric acid, and more rarely chromium and 
xirconium. In sdloys of aluminum other elements may be added to this list. 
The separation more commonly required is from iron, aluminum and iron 
being precipitated together as the hydroxides. In usual practice the two are 
weighed together as Fe20j and AUO3, after ignition to this form, and iron then 
determined, either on a separate portion of the sample, or by solution of the 
precipitate by fusion with sodium carbonate or potassium bisulphate and sub- 
sequent extraction with hydrochloric acid. The following procedures of sep- 
arations are given for special cases less commonly required in analysis. The 
chemist should be familiar with the substance with which he is working and 
have a general idea of its composition in order to be able to choose the 
correct procedure for estimation of the aluminum content. In ores and mate- 
rials to be used for production of aluminum compounds the results are reported 
in terms of the oxide-alumina, AI2O3, in alloys as the element, Al. 

Removal of Silica. This compound has already been considered under 
"Preparation and Solution of the Sample," Si02 being removed by taking the 
solution to dryness, dehydrating the oxide by additional heating in the oven, 
followed by extraction of the soluble constituents with dilute hydrochloric acid 
and filtration. Under the first procedure for solution of the ore by sulphuric and 
hydrofluoric acids silica is expelled as gaseous Sih\, 

Separation from Iron. 1. Aluminum hydroxide is precipitated by the 
addition of a salt of a weak acid to its neutral or slightly acid solution, iron 
remaining in solution. Details of the procedure for precipitation of aluminum 
hydroxide by means of sodium thiosulphate are given under "Gravimetric 
Methods for Determination of Aluminum," page 9. 

2. Aluminum chloride is precipitated from a concentrated solution of hydro- 
chloric acid and ether saturated with HCl gas. Details of the procedure are 
given under the gravimetric methods for aluminum, page 10. 

Note. The following additional procedures for separation of iron and alumina 
have been suggested: 

(a) Precipitation of iron as FeS in presence of organic acids, citric, tartaric, 
salicvlic, etc., aluminum remaining in solution. 

(6) Precipitating iron by adding sodium peroxide to a cold neutral solution of 
the elements until tne precipitate first formed dissolves, then decomposing the sodium 
ferrate by boiling, Fe(OH)s precipitates, Al remaining in solution. (Glaser, J. S. C. I., 
1897, 936.) 



6 ALUMINUM 

(c) The neutral solution of the elements is boiled with freshlv precipitated MnOi, 
which causes the precipitation of iron as Fe(0H)3, while aluminum remains in 
solution, (chromium also passes into the filtrate). 

(d) Precipitation of iron from acid solutions by means of amino-nitrosophenyl- 
hydroxylamine, (cupferron), aluminum remaining in solution. (O. Baudisch, Chem. 
Ztg., 83, 1298, 1905. Ibid., 85, 913, 1911; O. Baudisch and V. L. King, J. I. E. C, 
3,627,1911). 

(c) Precipitation of aluminum (together with phosphoric acid, if present), by 
phenylhydrazine, added to the reduced, weakly acid or neutral solutions. Iron, 
cobalt, nickel, calcium, and magnesium remain in solution. (Hess and Campbell. 
C. N., Ixxxi, 15S. Englea, J. S. C. I., 1898, 796.) 

(/) Electrolytic separation of iron by amalgamation with mercunr cathode and 
determining aluminum in the solution. (Kretzschmar, J. S. C. I., 1890, 1064; 
Kolin and Woodgate, J. S. C. I., 1889, 260.) 

Phosphoric Acid. In presence of phosphoric acid, the phosphates of iron 
and alumina together with the phosphates of the other elements of the group 
and those of the alkaline earths will be precipitated upon making the solution 
alkaline with ammonia. Should iron and alumina be the only elements of these 
two groups present in the solution, they may be precipitated together as phos- 
phates, iron determined by titration and calculated to the phosphate salt, 
and alumina obtained by difference. Occasionally, however, it is necessary to 
remove phosphoric acid. 

Removal of Phosphoric Acid. The material is fused with about six times 
its weight of a mixture of 4 parts NaaCOj and 1 part SiO« (silex), and the 
melt extracted with water containing ammonimn carbonate. Iron and aluminum 
remain on the filter, upon filtration, while sodium phosphate passes into solu- 
tion. Both the precipitate and filtrate contain silica. The precipitate of iron 
and alumina is dissolved in hydrochloric acid and taken to dryness, the 
residue dehydrated as usual, then treated with dilute hydrochloric acid and 
silica filtered off. The solution contains iron and aluminum in form of chlorides. 

Separation of Aluminum from Chromium. The solution is made strongly 
alkaline with sodium or potassium hydroxides and chromium oxidized by 
passing in chlorine gas or by adding bromine. The solution is now acidified 
with nitric acid and aluminum hydroxide precipitated by addition of ammo- 
nium hydroxide, chromium remaining in solution as a chromate. 

Separation of Aluminum from Manganese, Cobalt, Nickel, Zinc, the 
Alkaline Earths, and Alkalies. Iron and aluminum are precipitated as basic 
acetates, the other elements passing into solution. Details of the procedure 
are given under the basic acetate method on page 2(.0. 

In absence of phosphates, these elements do not interfere in the determi- 
nation of aluminum by precipitation as the hydroxide. 

Separation of Aluminum from Titanium. Details of the procedure are 
given under "Titanium." 

Separation of Aluminum from Uranium. Aluminum is precipitated as 
a carbonate in presence of a large amount of anunonium salts by addition of 
a large excess of ammonium carbonate and ammonium sulphide, while uranimn 
remains in solution as the complex compound UOi(COi)i(XH4)4. 

Separation from Glucinum. Aluminum is soluble in the fixed alkalies and 
remains in solution on boiling; glucinum also dissolves, but is precipitated on 
boiling. Glucinum is soluble in an excess of ammonium carbonate, aluminum 
is not. 

For additional separations see chapter on element in question. 



ALUMINUM 7 

GRAVIMETRIC METHODS FOR THE DETERMINATION 

OF ALUMINUM 

There are two general procedures for the gravimetric determination of 
aluminum. A. Direct determination, when it is possible to precipitate the 
hydroxide or phosphate of the element, free from impurities. B. Indirect deter- 
mination when the element is precipitated and weighed along with iron, the 
latter then determined by titration and aluminum estimated by difference. 

Deteritiination by Hydrolysis of an Aluminum Salt with Am- 
monium Hydroxide 

Principle. The method depends upon the hydrolysis of a soluble salt of alirnii- 
num by neutralizing the free and combined acid with ammonia. This hydrolysis 
takes place in presence of ammonium chloride, which prevents the precipitation 
of magnesium hydroxide by NH4OH, the common ion, NH4"^, repressing the 
ionization of the base, NH4OH. (See Notes.) The direct determination 
of aluminimi by this procedure excludes the presence of elements undergoing 
hydrolysis with similar conditions. Iron, chromiimi, titanium, zirconium, thal- 
lium, cerium interfere. In their presence a separation must be made. 

Reaction. AICI3+3NFI4OH =Al(OH),+3NH4Cl. 

If phosphoric acid is present in the solution aluminum will be precipitated 
as the phosphate, AIPO4. 

Procedure. To the solution, containing aluminum, free from phosphoric 
acid and the elements precipitated by ammonium hydroxide, are added 10 cc. 
of ammonium chloride (10%) and 5 cc. of concentrated nitric acid. The 
solution is diluted to about 150 cc. and heated to boiling. Upon cooling slightly, 
carbonate-free ammonium hydroxide is added slowly from a burette until a 
slight permanent precipitate forms, and then drop by drop until the solution 
reacts alkaline to litmus paper and the odor of ammonia is faintly perceptible. 
The precipitate is allowed to settle on the water bath for a few minutes, then 
filtered hot and washed first several times by decantation and finally on the 
filter with a hot solution of ammonium nitrate. (Twenty cc. strong nitric 
acid diluted and neutralized with ammonium hydroxide and made to 1000 cc.) 

The precipitate is purified, if other members of the ammonium sulphide 
group and following groups are present, as the gelatinous precipitate is apt to 
occlude some of these. This is accomplished by dissolving the precipitate in 
a small amount of hot, dilute hydrochloric acid, 1:1, the solution being caught 
in the beaker in which the first precipitation was made. The precipitation of 
the hydroxide is repeated exactly as is stated above. The precipitate, washed 
free of chlorides (AgNOi test), is drained of water and placed together with 
the filter paper in a platinum crucible. 

The ignition of the precipitate is conducted slowly at first until the paper 
is thoroughly charred, the heat is now increased to the full power of the Meker 
blast, the crucible being covered to prevent mechanical loss. Blasting for 
thirty minutes is generally sufficient to dehydrate the oxide, AUOi. It is advis- 
able, however, to repeat the heating until the weight becomes constant. The 
residue is weighed as AUOj. 

Al,O,X0.5303=Al. 



8 ALUMINUM 

Notes. Ammonia should be free from carbonates. Upon long ritanding with 
frequent exposure to air the ammonia takes up CO2, forming carbonate of ammonia. 
Freshly distilled an^nionia will be pure, the carbonate being precipitated by addition 
of lime in the distilling flask. Ammonia is best kept in a ceresine or paraffine 
bottle. It will then remain free from silica, which it invariably contains when con- 
fined in glass bottles. 

Long heating of the mixture containing the aluminum precipitate is objectionable. 

1. The solution is apt to become acid owing to the decomposition of ammonium 
salts and the volatilization of ammonia. 

2. The precipitate will become slimy and will be difficult to 'Wash and filter. 
It is preferable to redissolve and again precipitate if this condition occurs. 

3. The COt of the air is apt to be absorbed by the solution, causing the precip- 
itation of calcium carbonate, etc., should the solution be exposed for any length 
of time. 

4. Silica from the beaker will contaminate the precipitate. 

Hence it is advisable to filter as soon as possible after making the precipitation 
of Al(OH),. 

5. and S. No. 589, black band filter paper filters well and may be used to ad- 
vantage with precipitates of the nature of aluminum hydroxide. B. and A. No. B. 
filter is also good. 

Washing the precipitate with ammonium nitrate prevents the aluminum from 
passing through the filter and keeps it from packing, it favors the formation of 
the insoluble hydrogcl form of the hydrate wnile preventinj? the formation of the 
soluble hydrosol. Ammonium chloride may be used m place of nitrate.^ 

Alummum hydroxide is soluble in acids and alkalies. The ignited oxide, 
AlsOj, is insoluble in acetic acid but is soluble in mineral acids and the fixed alkalies, 
It is rendered very difficultly soluble in acids by strong ignition, generally requiring 
fusion with sodium carbonate or potassium bisulphate with subsequent acid treat- 
ment to efi'ect solution. 

AljO,. m.w,f 102.2; sv.ffr., 3.73 to 3.99; w.p., 2020° C. 

A yellow or reddish precipitate indicates the presence of iron, an element 
frequently present with aluminum. Should this be the case, iron must be deter- 
mined, either in a separate portion of the sample, or in the residue obtained b}r the 
procedure outlined. The amount of FcjOj is subtracted from the total residue, 
and AljOi obtained by difference. 

If phosphoric acid is present the phosphate of alumina will precipitate together 
with the phosphates of elements insoluble in alkaline solutions. Should phosphoric 
acid be present either its removal is essential, or the phosphate method for alumina 
should be followed. 

Fluorides hinder the precipitation of aluminum.* Evaporation to dryness and 
heating the residue to redness will transform fluorides to oxides and overcome this 
difficulty. 

Sulphates tends to hold up aluminuni from precipitation and a certain amount of 
sulphuric acid is occluded by the aluminum hydroxide precipitate. Magnesium is 
more apt to precipitate with alumina in presence of sulphates.' Ammonium chloride 
greatly lessens this difficulty. 

Traces of alumina may be recovered from the filtrate by evaporation to dryness, 
ignition and resolution with HCl. The A1(0H)3 is now precipitated with NH4OH. 

Since alumina absorbs moisture from the air, the crucible containing this compound 
should be kept covered in a desiccator until weighed. 

Ammonium hydroxide, in priesence of sufficient NH4CI, will not precipitate MgCOH)^, 
since the addition of NH4CI increases the ammonium ions in the solution and, by the 
common ion effect, represses the hydroxyl ions of the base, NH4OH, so that there are 
insufficient hydroxyl ions for the solubility product of Mg( OH)2 to be exceeded; there- 
fore magnesium remains in solution. A discus.sion of the theory of solubility product 
and law of mass action may be found in the author's work on Qualitative Cnemical 
Analysis, published by D. Van Nostrand Co. Reference is also made to Vol. I of The 
Elements of Quafitative Chemical Analysis, by Julius Stieglitz, publ. by the Century Co. 

» W. Blum, Jour. Am. Chem. Soc, 38, 7, 1282, 1916. C. F. Sidener and Earl Petti- 
john, Jour. Ind. Eng. Chem., 8, 8, 714, Aug., 1916. 

« E. P. Veitch, Jour. Am. Chem. Soc, 22, 246, 1900. W. R. Bloor, ibid., 20, 16a3, 
1907. L. P. Curtman and H. Dubin, ibid., S4, 1485, 1912. 



ALUMINUM 9 

Determination of Aluminum by Hydrolysis, Neutralizing the 
Mfneral Acid by Addition of a Salt of a Weak Acid. 
Sodium Thiosulphate Method ^ 

If a salt of a weak acid and strong base is added to a neutral or slightly acid 
solution of a^^ aluminum salt containing a mineral acid, transposition takes 
place and aluLdinum is hydrolyzcd. 

Reaction. 2AlCl,+3Na2S,0,+3H,0 =2Al(OH)3+6NaCl+3SO,-|-3S. 

Procedure. If the solution is acid, dilute ammonia is added until a pre- 
cipitate forms that dissolves with difficulty, but not enough ammonium hydroxide 
to cause a permanent precipitation. The solution is diluted so that it con- 
tains about O.l g. Al per 200 cc, then an excess of sodium thiosulphate is added, 
and the solution is boiled free of SOj. A1(0H)3 precipitates along with free 
sulphur. If iron is absent it is advisable to add a few drops of ammonium 
hydroxide until the solution has a slight odor of ammonia. The mixture again 
boiled is filtered and the residue of Al(OH)j and sulphur washed with hot water 
containing ammoniimi chloride or nitrate. The precipitate is dried, separated 
from the filter, the latter ignited and the ash added to the main precipitate. 
Alumina is now determined by blasting to constant weight, the residue being 
weighed as AUOa. 

Notes. The above method may be employed for separation of alununiim from 
iron, the addition of ammonia, following the neutralization of the mineral acid by 
thiosulphate being omitted. The precipitation of A1(0H)3 by this procedure gives 
a more dense and better filtering precipitate than does ammonia alone. 

Note. G. Wynkoop suggests the use of sodium nitrite as the salt of a weak acid 
for neutralizing the mineral acid.* 

Reaction. 2AlCl3H-6HOH=2Al(OH),+6HCl and 

6HCH-6NaNO, = 6NaCl -hSHaO -1-3NO +3N0,. 

Precipitation of Aluminum as a Phosphate 

Principle. This procedure, developed by Camot,* is of special value in 
determination of aluminum in iron and steel. It is founded on the reaction that 
aluminum is precipitated as the neutral phosphate, from a boiling solution faintly 
acid with acetic acid. Iron, reduced to the ferrous condition by addition of 
sodium hyposulphite, does not interfere. 

Procedure. A sample of 10 grams of iron or steel, in a platinum dish, covered 
with a piece of platinum foil, is dissolved by addition of hydrochloric acid. 
The solution is diluted to about 100 cc. and filtered into a flask, the residue of 
carbon, silica, etc., is washed thoroughly and the filtrate is neutralized by 
addition of ammonium hydroxide and ammonium carbonate; no permanent 
precipitate should form. A little sodium hyposulphite is added, and when 
the liquid, at first violet, becomes colorless, 2 or 3 cc. of a saturated solu- 

» Method by Chancel, Compt. rend. (1858), 46, 987. 

« J. Am. Ch. Soc, 19, 434 (1897). The method may be found in Treadwell and 
Hall " Quantitative Analysis/' 4th ed., p. 8-5. 

'A. Camot, Moniteur Scientifique, 1891, p. 14. 



10 ALUMINUM 

tion of sodium phosphate and 5 or 6 grams of sodium acetate, dissolved in a 
little water, are added. The solution is boiled until free of sulphurous acid odor 
(about three-quarters of an hour). The solution is filtered off from the pre- 
cipitated aluminum phosphate (mixed with a little silica and ferric phosphate) 
and washed with boiling water. The precipitate on the filter is treated with hot 
dilute hydrochloric acid the filtrate caught in a platinum dish, and then evap- 
orated to dryness and heated at 110° C. for an hour to dehydrate silica. The 
residue is taken up with dilute hydrochloric acid and the solution filtered free 
of silica. Upon dilution to about 100 cc. with cold water, the solution is neu- 
tralized as before, a little hyposulphite is added to the cold solution and then 
a mixture of 2 grams of sodium hyposulphite and 2 grams of sodium acetate. 
The material is boiled for half an hour or more, filtered and the aluminum phos- 
phate residue washed with hot water, then dried, ignited and weighed as aluminum 
phosphate. The residue contains 22.19% Al. 

AIPO4XO.2219 = Al. AIPO4XO.4I85 = AUG.. 

Note. Properties of AIPO4, m.tr., 122.14; sp.^., 2.59; infusible, insoluble, iu IlfO 
and in HCaHsOz, soluble in mineral acids and in alkalies; white, amorphous salt.. 

Precipitation of Aluminum as Aluminum Chloride^ 

Principle. Gooch and Havens found that aluminum chloride is practically 
insoluble in a mixture of concentrated hydrochloric acid and ether saturated with 
HCl gas, 5 parts of AICI1.6H2G equivalent to 1 part of AUGi dissolving in 125,000 
parts of the mixture. The method serves for a separation of aluminum from 
iron, berillium, zinc, copper, mercury and bismuth, the chlorides of these ele- 
ments being soluble under the above conditions. Barium, however, is precipi- 
tated as a chloride with aluminum, if it is present in the solution. 

Procedure. To the concentrated aqueous solution of aluminum is added 
a convenient volume of strong hydrochloric acid (15 to 25 cc.) and an equal 
volume of ether. The mixture is best placed in a large platinum crucible, 
which is kept cool in running water. HCl gas is passed into the solution to 
saturation. The precipitated chloride of aluminum is filtered upon asbestos 
in a weighed Gooch crucible and then washed with a mixture of ether and water 
1:1, saturated with HCl gas. The precipitate is dried for half an hour at 150® 
C, then covered with a layer of C.P. mercuric oxide (1 gram) and heated at first, 
gently over a low fiame (hood) and then blasted to constant weight. The 
residue is weighed as AlaOs. 

NoTErt. HCl gas is generated by dropping strong sulphuric acid into concen- 
trated hydrochloric acid according to the procedure described under the determina- 
tion of arsenic by volatilization as arsenious chloride. The gas may be produced 
in a Kipp generator by the action of concentrated sulphuric acid on ammonium 
chloride. 

The filtrate from aluminum contains iron, berillium, copper, zinc, etc., if these 
are present in the original solution. If much iron is present it is necessary to increase 
the amount of ether to prevent precipitation of the ferric salt. 

' F. A. Gooch and F. S. Havens, Am. Jour. Sci. (4), 11, 416. F. A. Gooch " Methods 
in Chemical Analysis." 



ALUMINUM 11 



VOLUMETRIC METHODS FOR THE DETERMINATION OF 

ALUMINUM 

Volumetric Determination of Combined Alumina in Aluminum 

Sulphate and Aluminum Salts 

Introduction. Aluminum salts dissociate in hot solutions and react acid to 
phenolphthalein indicator; the acid readily combines with fixed alkalies, forming 
the neutral alkali salt. The end point of the reaction is indicated by the pink 
color produced upon phenolphthalein by the excess of alkali. From the amount 
of caustic required the percentage of combined AUOi may be calculated. The 
following reaction takes place: 

Al,(S04),+6NaOH =2Al(OH),+3Na,S04. 

Procedure. The factor weight,* 3.4067 grams, is dissolved in a 4-in. casserole 
with 100 cc. of distilled water, 1 cc. of phenolphthalein indicator added, and 
the sample titrated boiling hot * with N/2 NaOH, added from a chamber burette, 
graduated from 50 to 100 cc. in tenths of a cc* The solution is kept boiling 
during the titration and is constantly stirred. Towards the end of the reaction 
the alkali is added cautiously drop by drop until a permanent pink color is 
obtained. 

Cc. of NaOH required divided by 4 =p>er cent combined AUOi.« 

Combined AUOi+free AlsOs = total AljOa. 

Notes. If iron is present a correction must be made for it after determining the 
ferrous and ferric forms as eiven below. 

The amount of phenolphthalein indicator used should be the same in each deter- 
mination. An excess of indicator causes low results. It has been noted in case of 
alums where iron does not interfere that best results are obtained with three or four 
drops of phenolphthalein solution. Iron tends to mask the end point, hence a larger 
amount of indicator is necessary if this is present. 

Correction for Iron if Present. Since iron salts will also dissociate and 
titrate with aluminum salts, by this method a correction has to be made for iron 
if present. Total AlaOi in presence of iron = 

combined AljO,-(FeOX.47+Fe20tX.64)+basic AUOt+an additive factor. 

The additive factor is obtained by subtracting 

(Combined AUOt+basic AlaOi) -(FeOX.47+Fe,0,X.64) volumetric, 

from total AljOs obtained by gravimetric analysis of an average sample. 



* Large samples must be taken for salts containing less than 13 per cent AUd if 
the chamber burette is to be used. E.g., potash alum twice this amount is advisable. 

« Otto Schmatolla, Berichte, xxxviii. No. 4. C. N., 91-2375-236 (1905). 

'If free acid is present (see next method), the equivalent volume in terms of i N 
acid must be deducted from the total titration for combined alumina before divi^ng 
by 4. 



12 ALUMINUM 

Ferrous Iron, Ferric Iron, and Total Iron. A five-gram sample is dis- 
solved in water and the iron oxidized with a few drops of strong potassium 
permanganate solution; the solution should be pink; the excess of permanganate 
is destroyed by a drop or so of normal oxalic acid solution and the total iron 
determined by stannous chloride solution method for iron. On a separate 
sample ferric iron is determined. Ten grams of the sample are dissolved in an 
Erlenmeyer flask by boiling with hydrochloric acid, 2 : 1, in an atmosphere of 
COf to prevent oxidation, and the iron titrated with standard stannous chloride. 
The difference between total iron as Fe^Oj and ferric oxide = ferrous iron in terms 
of Fe,Oi. This multiplied by .9 = FeO. 

Combined Sulphuric Acid 

Provided no free acid is present, the per cent combined sulphuric acid in 
aluminum sulphate is obtained by multiplying the cc. caustic titration for total 
alumina by 0.72. 

In case free acid is present, the per cent free acid deducted from total acid 
found by titration gives combined acid. 

Sulphuric acid combined with the fixed alkalies is not titrated. 

Determination of Free Alumina or Free Acid by the Potassium 

Fluoride Method 

Introduction. The method suggested by T. J. I. Craig (J. S. C. I. 80, 
185), has been modified by the author,* after a personal investigation of the 
details involved. In this modified form it has been used successfully as a rapid 
works method. Frequent gravimetric checks on a large number of determi- 
nations have shown it to be accurate. 

The procedure is based upon the fact that an excess of neutral pK)tassium 
fluoride decomposes aluminum salts, forming two stable compounds, which react 
neutral to phenolphthalein, . while the free acid remains unaltered, the fol- 
lowing reaction taking place: 

Ala(S04),+12KF+xH^04=2AlF,3KF+3K2S04+xH,S04. 

The precipitate A1F»3KF is insoluble in an excess of the potassium fluoride 
reagent and is not appreciably attacked by acids or alkalies. Although 
theoretically about 7 parts by weight of potassium fluoride is sufficient to com- 
bine with 1 part of aluminum sulphate, in practice it is advisable to use twice 
this amount. 

Reagents Re<iuired. Half Normal solutions of sulphuric acid and potassium 
hydroxide, (sodium h^ydroxide may be used.) 

Phenolphthalein indicator, 0.1% alcoholic solution. 

Potassium fluoride solution; made by dissolving 1000 grams of potassium fluoride 
in about 1200 cc. of hot, COj-free water, then neutralizing the solution with hydro- 
fluoric acid or |>otassium hydroxide as the reagent may require, using 5 cc. of phenol- 
phthalein as indicator. Dilute sulphuric acid may be used in place of hydrofluoric 
acid in the final acid adjustment to get a neutral product. Onecc. of the solution 
in 10 cc. of COr-free wat<»r should appear a faint pmk. The concentrated mix is 
filtered if necessary and then diluted to 2000 cc. with COj-fn^ water. The gravity 
will now be approximately 1..32 or about iJ6'' B6, One cc. contains 0.5 g. potassium 
fluoride. 

1 W. W. Scott. 



ALUMINUM 13 



Method of Procedure 

Solids. 3.4067 g. of the finely ground sample, or an equivalent amount in 
solution (100 cc. of sample containing 34.067 g. per liter), are taken for analysis 
The powder is dissolved by boiling with 100 cc. of distilled water in a4-in- 
casserole with clock glass cover. To the hot solution 10 cc. of N/2 H2SO4 are added, 
and after coolmg to room temperature, 20° C, 18 to 20 cc. of the potassium 
fluoride reagent are added and 0.5 cc. of phenolphthalein. The solution is now 
titrated with N/2 KOH, added drop by drop until a delicate pink color, per- 
sisting for one minute, is obtained. This titration shows whether the product 
is basic or acid. 

BoBic Alumina. This is indicated when the alkali back-titration is less 
than the amount of acid added. Free AUOj = (cc. H2SO4 — cc. KOH) -^4. 

Free Acid. In case the back-titration of the alkali is greater than the cc. of 
acid added, free acid is present. Free acid = (cc. KOH — cc. HjSO*) X0.72. 

Liquors. In works control it is necessary to test the concentrated liquors 
to ascertain whether these are basic or acidic. The B6. or sp.gr. of the solu- 
tion having been taken, 5 cc. is diluted to 100 cc. with distilled, COs-free water. 
If HjS is present, it is expelled by boiling the solution, which should be acid, 10 
cc. of N/2 H1SO4 is added, the solution cooled, and KF and phenolphthalein 
added and the titration made as in case of solids. 

If basic (cc. H2SO4-CC. KOH) X (.0245 X. 3473 X 100) -^ (5 Xsp.gr.) ^Al^O,. 

If acid (cc. KOH — cc. H2SO4X2.45) -^ wt. of sample =per cent free acid (H2SO4). 

If neutral, the back titration of the alkali is the same as the cc. acid 
added. 

Notes. COj-free water must always be used when phenolphthalein indicator is 
necessary. This may be obtained b^ boiling distilled water for several minutes to 
expel COj. This reagent is very sensitive to carbonic acid. 

If the sample does not dissolve clear, a prolonged digestion with previous a^idi- 
tion of the required amount of standard acid, 10 cc, is advisable. This is best 
accomplished in an Earlnmeyer flask with a return condenser. 

Darkening of the solution during the back titration with the alkali, indicates that 
an insufficient amount of fluoride has been added. If this is the case it will be necessary 
to make a fresh determination. 

The fluoride method has the following advantages. Determinations may be made 
by gas or electric light. The end point is easily detected . No neutral standard is nec- 
essary as in case of the tint method. 

Ammonium salts, if present, must be expelled by boiling the sample with an excess 
of standard KOH and this excess determined. 

3.4067 = 2.45225 X .3473 X 4 (i .e. gms. H28O4 per 100 cc. N/2 acid multiplied by 4 times 
factor to equivalent AljOa). Derived directly from mol. wt. of AljOj^ (.1022X100X4) 
-^(6X2). 0.72 = 2.8792-^4 (i.e. factor AI2O3 to H2S04H-4). 

The main details of the above volumetric procedures were worked out at the Laurel 
Hill Laboratory, General Chemical Company, and are published by courtesy of this 
company. 

The author is indebted to Mr. W. S. Allen for his criticismand valuable suggestions 
in the volumetric procedures for determining alumina.] 



14 ALUMINUM 

Detection and Colorimetric Estimation of Minute Amounts of 
Aluminum with Alizarin S.— Atack's Method ^ 

The reagent used is a 0.1% filtered solution of conunercial alizarin S, the 
sodium salt of alizarin monosulphonic acid (yellow with acids, purple with 
alkalies). 

Test To 5 cc. of the neutral or acid solution under examination is added 
1 cc. of the reagent, and then ammonia until the solution is alkaline, as shown 
by the purple color. The solution is boiled for a few moments, allowed to 
cool, and then acidified with dilute acetic acid, when red coloration or pre- 
cipitate remaining is conclusive evidence of the presence of aluminum. The 
red calcium, strontium, barium, zinc and magnesium salts, and salts of other 
metals later than Group II are readily soluble in cold dilute acetic acid, and do 
not interefere with the coloration. 

Phosphates or chromium do not interfere and comparatively large amounts 
of iron may be present (0.003 milligram Al in presence of 1 milligram ferric iron, 
10 milligrams chromiiun salt). In presence of greater quantities of iron citric 
acid is added to keep this in solution. 

Delicacy of the Test One part of aluminiun may be detected in 10 million 
parts of water. 

Quantitative Estimation, Colorimetric 

Procedure. The original solution (5 to 20 cc.) is acidified with hydro- 
chloric or sulphuric acid. Ten cc. of glycerin and 5 cc. of a 1% solution of 
alizarin S are added, the solution made up to about 40 cc. with water (in pres- 
ence of much iron or chromium citric acid is added to form the double citrates) 
and then rendered slightly ammoniacal. After standing for five minutes, the 
cold solution is acidified with dilute acetic acid, the alizarin S acting as indicator 
(red coloration) until no further change in the coloration occurs. The hquid 
is then made up to 50 cc. and compared with a standard. Suitable amounts 
of aluminum for estimation are 0.005 to 0.05 milligrams, the solution under 
examination being suitably diluted if necessary. 

BAUXITE ANALYSIS 2 

Characteristic bauxites H,0 SiOj FesOt AliOi TiO, 

Arkansas 6.4% 1.43% 87.3% 3.99% 

Georgia 36% 9-15 1-14 42-62 1.8-2.3 

Tennessee 27.6 18.4 4.1 49.9 

Sampling. The bauxite received in cars is sampled during the unloading 
according to the standard procedure for ores. If the sample is a composite 
aliquot parts of the total weights are taken and mixed, e.g., suppose three cars 
contained respectively 23,000, 32,500, and 26,340 pounds, then the aliquots 

1 F. W. Atack,Jour.Soc.Chem.Ind.; 34,936(1915); C. A.9;23; 3186 (1915). 

' Bauxite is the only ore of aluminum of commercial importance. Pure alumina, 
corundum, is too valuable for commercial use. Clay, the most abundant of alumina- 
bearing substances, may eventually be used as a source for aluminum, but, by the 
present methods of extraction, the alumina from clay is not commercially available. 



ALUMINUM 15 

would be 23, 32.5 and 26.34 pounds, which mixed, would make a representa- 
tive sample of the shipment. The ore is broken down, quartered, ground down 
and again quartered. The moisture is determined on 1000 grams, dried in 
the oven at 100** C. for one hour, the sample being spread out on a sheet of 
manilla paper. The dried sample is placed in a large bottle for analysis. 

Procedure for Evaluation of the Ore. A method for obtaining in solution 
the available alumina and soluble constituents of bauxite has been given under 
Preparation and Solution of the Sample. 

Insoluble Residue. The residue on the filter paper is ignited in a plat- 
inum dish over a low fiame until the paper chars, and then over a good Meker 
blast for 15 to 20 minutes. 

Weight of the residue X 20 =per cent insoluble residue. 

Soluble Alumina. 100 cc. of the above solution (0.5 g.) is diluted with an 
equal volume of water, 10 cc. of hydrochloric and 2 cc. of nitric acids added 
and the solution boiled. Iron and alumina are now precipitated and deter- 
mined in the usual way. 

Soluble Iron. 200 cc. of the solution (1.0 g.), is oxidized by adding a 
few crystals of potassium chlorate and the solution taken to dryness. The 
residue is taken up with 10 to 15 cc. of concentrated hydrochloric acid and again 
evaporated to dryness to expel chlorine. Then taken up with 25 cc. hydro- 
chloric acid and the iron determined by titration. The stannous chloride method 
is used for samples containing less than 5% iron and the dichromate method 
for ores containing over 5%. 

Determination of Total Silica, Titanium Oxide, Ferric Oxide and Alumina 

The method by the Aluminum Company of America is to digest 1 gram of 
the dried bauxite in 90 cc. of an acid mixture containing 12 parts of dilute sul- 
phuric acid, 1 : 3, together with 6 parts of strong hydrochloric acid and 2 parts 
of nitric by volume, to this are added 10 cc. of concentrated sulphuric acid. 
The mixture is heated until sulphuric acid fumes are evolved, then diluted with 
water and filtered. 

Silica. The residue is ignited and the ash fused with potassium bisulphate. 
The cooled fusion is taken up with 5 cc. sulphuric acid and 20 cc. of water and 
digested until only a white residue remains. This filtered off, washed and 
ignited -SiOs. 

Titanium Oxide. This is best determined colorometrically on a 0.1 gram 
sample according to the procedure outlined in the chapter on Titanium. 

Iron and Alumina. These are determined by the usual procedure; — oxida- 
tion with potassium chlorate, precipitation with ammonium hydroxide and 
ignition. Iron may be determined in a separate sample (100 cc. =0.5 g.) 
by titration. AlaOj = difference between weighed oxides and Fead, after sub- 
tracting TiOf if present. 



16 ALUMINUM 



DETERMINATION OF ALUMINUM IN IRON AND STEEL i 

The method is especially adapted for determination of aluminum in iron and 
steel, but may be extended to iron ores and materials high in iron. 

Procedure. Solution. Ten grams of iron or steel are dissolved by adding 
about 50 CO. of hot hydrochloric acid, 1:1, preferably in a platinum dish, covered 
with a platiniun foil. 

Precipitation. When the solution of iron is complete, it is diluted to 
about 100 cc. and filtered free of carbon, silica, etc. Two grams of sodium phos- 
phate are added and the solution neutralized with ammonimn hydroxide or 
carbonate, then cleared by hydrochloric acid with about 1 cc. excess. Twenty 
cc. of acetic acid are now added and the solution diluted to 300 to 400 cc. with 
hot water and, on boiling, 10 grams of sodium thiosulphate added. The solu- 
tion is boiled free of sulphurous acid, (no odor of SO2) about 20 to 30 minutes 
being necessary. The phosphate is filtered off and washed with hot water. It 
is again dissolved in a little hydrochloric acid and aluminum reprecipitated by 
neutralizing with ammonium hydroxide and adding about 1 gram of sodium 
phosphate together with 10 grams of sodium thiosulphate, following the above 
procedure. The precipitate will now be free of iron. 

Ignition and Calculation. The precipitate and filter are ignited wet, 
first over a low flame, then gradually increasing the heat to full blast of a 
Meker burner. The residue contains 22.19% Al or 41.85% of AlaOi. 

Factor AIPO4 to Al = .2219. 

Factor AIPO4 to AI2O, = .4185. 

Notes. Interfering substances. Copper may be removed by HaS. Other mem- 
bers of this group will also be eliiiiinated. 

Manganese and nickel are eliminated together with small amounts of iron at the 
second precipitation. 

Titanium may be estimation colorimetrically or separated from alumina. 

Vanadium, if present, may be separated according to directions given in the chapter 
on Vanadium. 

Chromium is eliminated by fusion of the mixed phosphates with NajCOj, extrac- 
tion with water, and precipitation of aluminum phosphate by adding ammonium 
acetate and sodium phosphate. Chromium remains in solution. 

ANALYSIS OF METALLIC ALUMINUM 2 

Determination of Silicon 

Acid Mixture: 400 cc. cone, nitric acid. 1200 cc. cone, hydrochloric acid. 

600 cc. cone, sulphuric acid. 1800 cc. water. 

Fusion Method 

Dissolve 1 gram of well mixed drillings in 35 cc. of acid mixture using a 
4^inch porcelain dish with a 5-inch cover glass. When the drillings are com- 
pletely dissolved, evaporate the solution not only to fuming but to complete 

* Arnold and Ibbotson, "Steel Works Materials." Stillman, ^'Engineering Chem- 
istry ." "A Rapid Method for the Determination of Aluminum in iron and Steel," 
C. N., 61, 313. "On the Determination of Minute Quantities of Al in Iron and 
Steel,'^ J. E. Stead, J. S. C. I., 1889, 956. 

•Standard Method of Analysis of the Aluminum Company of America. By 
courtesy of Mr. E.* Blough, Chief Chemist. 



ALUMINUM 17 

diyness, and bake. This insures the freedom of the solution from hydrochloric 
and nitric acids, and the complete dehydration of the silica. Take up the 
residue with 10 cc. 25 per cent sulphuric acid and about 100 cc. of water; boU 
to complete solution of the sulphate, filter, wash well and ignite. Fuse the 
residue with eight to ten times its weight of sodium carbonate and take up the 
fused mass in a porcelain dish with sulphuric acid (1 : 1). Evaporate the 
resulting solution until copious fumes are evolved, which will cause the separa- 
tion of the silica; dilute carefully, boil, filter, wash well and ignite in a platinum 
crucible and weigh. Treat the ash with hydrofluoric acid and a few drops of 
sulphuric acid; carefully ignite and weigh. The difference in the two weights 
obtained above represents the silicon as silica. 

Calculate the silica to silicon by the factor 0.4693. 

Graphitic Silicon 

Aluminum, sometimes if not always, contains some silicon in the graphitic 
state; this graphitic silicon does not oxidize to SiOt on ignition and is not 
volatile with HF, which two characteristics distinguish it from amorphous silicon. 

To determine graphitic silicon the mixture of Si and SiOt obtained as in the 
solution method is treated in a weighed platinum crucible with 2-3 drops of 
HiS04 and 2-3 cc. HF. 

The brown residue of Si remaining is strongly ignited and weighed; the 
silicon remaining is that which was in the metal in the graphitic state. 

Determination of Iron 

Permanganate Method 

Cool the filtrate obtained from solution of the sample in acid mixture 
(see page 16) and reduce the iron present by passing the solution through a Jones 
reductor. Titrate immediately with a solution of potassium permanganate of 
such strength that 1 cc. equals 0.0010 gram iron. 

In all cases the precautions given for use of the Jones reductor should be 
observed, and explicit directions given in the chapter on Iron, carefully followed. 
A blank determination is made by carrying out a regular iron determination 
with the metal sample omitted. The amount of potassium permaiiganate re- 
quired to give the blank a distinct color is subtracted from the amount re- 
quired to give the same color to each reduced solution. 

The author acknowledges his indebtedness to Mr. W. S. Allen, Mr. J. P. Kelly and 
Dr. F. E. Hale for review and criticism of the subject. 



ANTIMONY 

Wilfred W. Scott 
Slh at.wt. 1S0.2; sp.gr. 6.62^; m.p. 630^C>; b.p. 1440^0^; oxides, SbtOu 

DETECTION 

Hydrogen Sulphide precipitates the orange-colored sulphide of antimony 
from fairly strong hydrochloric acid solutions (1 : 4) in which several mem- 
bers of the group remain dissolved. Arsenic is also precipitated. The latter 
may be removed by boiling the solution containing the trichloride, AsCU being 
volatile. 

If antimony is already present as a sulphide, together with other elements 
of the hydrogen sulphide group, it may be dissolved out by treating the 
precipitate with sodium hydroxide, potassium hydroxide, sodium sulphide, 
ammonium polysulphide in solution. Antimony sulphide is reprecipitated 
upon acidifying the filtrate. Arsenic and tin will also be precipitated with 
antimony if they are present in the original precipitate. Should a separation 
be necessary, the precipitate is dissolved with hot concentrated hydrochloric 
acid, with the addition of crystals of potassium chlorate, from time to time, 
until the sulphides dissolve. The solution is placed in a Marsh apparatus, pure 
zinc added and the evolved gases passed into a neutral solution of silver nitrate. 
The black precipitate of silver antimonide and metallic silver are filtered off, 
washed free of arsenous acid, and the antimonide dissolved in strong hydro- 
chloric acid (silver remains insoluble). The orange-colored antimony sulphide 
may now be precipitated by diluting the solution with water and passing in 
H2S gas to saturation. 

Minerals which contain antimony, when heated alone or with 3 to 4 parts 
of fusion mixture (K2C08 and Na^COs), on charcoal, yield dense white fumes, 
a portion of the oxide remaining as a white incrustation on the charcoal. A 
drop of ammonium sulphide placed upon this sublimate gives a deep orange stain. 

Hydrolysis. Most of the inorganic antimony salts are decomposed by 
water, forming insoluble basic salts, which in turn break down to the oxide of 
antimony and free acid. An excess of tartaric acid prevents this precipitation. 

Traces of Antimony. Nascent hydrogen liberated by the action of zinc 
and hydrochloric or sulphuric acid reacts upon antimony compounds with the 
formation of stibine. This gas produces a black stain on mercuric chloride or 
silver nitrate paper. Details of the procedure are given under the quantita- 
tive method for determining minute amounts of antimony. 

Distinction between Antimonous and Antimonic Salts. 

Chromates form with antimonous salts green chromic salts and antimonic salts. 

PotoBsiutn Iodide reduces antimonic salts, free iodine being liberated. 

* Van Nostrand's Chem. Annual, Olsen, 3d Ed. 
« Cir. 35, U. S. Bureau of Standards. 

18 



ANTIMONY 19 



ESTIMATION 

The determination of antimony is required in the evaluation of antimony 
ores — stibnite, SbjSi; valentinite, SbjOa, etc. It is generally required in the 
complete analysis of minerals of nickel, lead, copper, silver, in which antimony 
generally occurs as a sulphide. The determination is required in the analysis 
of Britannia metal, bearing and antifriction metals, type metal and hard lead; 
in the analysis of certain mordants, antimony salts, vulcanized rubber, etc. It 
is looked for as an undesirable impurity in certain food products. 

Preparation and Solution of the Sample 

In dissolving the substance containing antimony it must be remembered 
that metallic antimony is practically insoluble in cold dilute hydrochloric, nitric 
or sulphuric acid and the oxides, Sb^Os or Sb206, are precipitated in strong nitric 
acid. The element, however, is readily soluble in hydrochloric acid contain- 
ing an oxidizing agent, such as nitric acid, potassium chlorate, chlorine, bromine, 
etc. The oxides of antimony are soluble in hydrochloric acid and the caustic 
alkalies. 

Solution of Sulphide Ores, Low-grade Oxides, etc.^ 

0.5 to 1 gram of the finely ground ore, placed in a Kjeldahl flask, is 
mixed with 5 to 7 grams of anmionium sulphate, 1 gram of potassium sul- 
phate, and 10 cc. of strong sulphuric acid. About 0.5 gram of tartaric acid, 
or a piece of filter paper, is added to reduce arsenic and antimony and the mixture 
heated, gradually at first, and then with the full Bunsen flame. The heating 
is continued until the carbon is completely oxidized and most of the free acid 
driven oflF, leaving a clean fusion from which anmionium sulphate is volatilizing. 
The melt is now cooled over the bottom and sides of the flask by gently rotating 
during the cooling. 

About 50 cc. of dilute hydrochloric acid (1:1) are added and the melt dis- 
solved by warming gently. The contents of the Kjeldahl flask are transferred 
to an Erienmeyer flask, the Kjeldahl being rinsed out with 25 cc. of strong 
hydrochloric acid. Arsenic sulphide may now be precipitated with HjS from 
the strongly acid solution, whereas antimony, etc., remain in solution. The 
sulphide is filtered off through a double filter, that has been moistened with 
hydrochloric acid (2:1), a platinum cone supporting the filter to prevent its 
breaking. The flask is rinsed out with hydrochloric acid (2 : 1). The pre- 
cipitate is washed at least six times with the acid. Antimony passes into the 
filtrate together with other elements of the ore. 

The filtrate is diluted with double its volume of warm water and then is 
saturated with hydrogen sulphide. Antimony sulphide, together with other 
elements of the Hydrogen Sulphide Group, will precipitate. These are washed 
with hydrogen sulphide water. Antimony sulphide may now be dissolved by 
addition of sodium sulphide and caustic solution (separation from Cu, Pb, Cd, 
Bi, etc.) (5 to 10 cc. of a mix of 60 grams NasS with 40 grams of NaOH 
diluted to 1000 cc). 

^ Method of A. H. Low modified. 



20 ANTIMONY 

The solution containing the antimony is treated with about 2 grams of 
potassium sulphate and 10 cc. of strong sulphuric acid and heated as before, 
to destroy liberated sulphur and expel most of the free acid. The melt is dis- 
solved in hydrochloric acid, and the antimony titrated according to one of the 
volumetric procedures given under ** Volumetric Methods." 

Note. An insoluble residue remaining from the acid extraction of the first melt 
may be dissolved by fusion with sodium hvdroxide and extraction of the melt with hot 
water. If a precipitate forms when this alkaline solution is acidified with hydrochloric 
acid, the presence of barium sulphate is indicated. 

Decomposition of the Ores by Fusion with Sodium Hydroxide. 

Oxides. 0.5 to 1 gram of the powdered ore is mixed with about 10 
grams of sodium hydroxide and placed in a thin-walled iron crucible of 60 cc. 
capacity. It is advisable to fuse a portion of the alkali hydroxide in the cru- 
cible with a pinch of potassium nitrate and then add the ore mixed with the 
remainder of the sodium hydroxide. The covered crucible is heated until the 
fusion becomes homogeneous. The melt is poured out on a large nickel crucible 
cover or shallow dish. On cooling, the cake is detached and placed in a cas- 
serole containing water, any adhering cake on the cover, or melt remaining in 
the iron crucible, being dissolved with dilute hydrochloric acid and added to 
the sample in the casserole. About 30 to 40 cc. of strong hydrochloric acid are 
now added and the mixture heated (casserole covered) until the melt has dis- 
solved. Two to 3 grams of tartaric acid having been added to keep anti- 
mony dissolved, the solution is diluted to about 300 cc, and antimony is then 
precipitated as the sulphide with hydrogen sulphide. The treatment of the 
precipitate at this stage has been given in the '* Solution of Sulphide Ores." 

Sulphides. Howard and Harrison * recommend the following procedure 
for fusion of sulphide ores with caustic: 0.5 gram of the powdered ore is fused 
with a mixture of 8 grams of sodium carbonate and sodium peroxide, 1:1, 
in a nickel crucible. The cooled melt is dissolved with sufficient hydrochloric 
acid to neutralize the alkali and about 15 cc. of strong acid added in excess. 
The solution is diluted to 250 cc, antimony being kept in solution by addition 
of potassium chlorate. An aliquot portion of the solution is taken, antimony 
reduced by metabisulphite and titrated ^ith iodine. 

Treatment of Speisses, Slags, Mattes, etc.* 0.5 to 2 grams of the 
sample is treated with 10 to 15 cc of strong nitric acid and the mixture taken 
to dryness. Fifteen cc. of strong hydrochloric acid are added and the sample 
transferred to a 350-cc. flask, additional hydrochloric acid being used to wash 
out the beaker. Arsenic is precipitated from the strong acid solution as the 
sulphide, and antimony determined in the filtrate. 

Solution of AllojTS. Alloys are generally decomposed by treatment with 
mixtures of hydrochloric acid together with an oxidizing agent — nitric acid, 
potassium chlorate, bromine, etc. The subject is taken up in detail in the chapter 
on alloys. 

The alloy drillings are treated with strong hydrochloric acid, a little bro- 
mine added, and the mixture heated until the alloy dissolves, additional bromine 
being added from time to time if necesvsary. The excess bromine is removed 
by heating gently to boiling. The higher oxides are reduced by addition of 

» Phar. Jour., 1909, 88, 147. 
* H. E. Hooper's method. 



ANTIMONY 21 

sodium metabisulphite and the sulphides precipitated, as usual, with hydrogen 
sulphide. Arsenic may now be volatilized by boiling, and antimony titrated 
with iodine or potassium bromate. 

Alloys of Antimony, Lead and Tin. 0.5 to 1 gram of the finely divided 
alloy is warmed with 100 cc. of strong hydrochloric acid until the action sub- 
sides. Solid iodine is now added, in small quantities at a time, until the alloy 
completely dissolves. The excess of iodine is now removed by boiling and the 
small amount of free iodine remaining neutralized with a few drops of a weak 
solution of sodium thiosulphate. Although tin is oxidized to the higher state, 
antimony is not oxidized by iodine in acid solution beyond the trivalent form. 
The solution may now be titrated with standard iodine in presence of an excess 
of sodium bicarbonate according to the procedure given under the volumetric 
methods. 

Hard Lead. The method of solution and titration are given under * Potas- 
sium Bromate Method for Determining Antimony." 

Antimony in Rubber Goods.' Three grams of the finely rasped rubber 
are treated in a Kjeldahl flask with 40 to 45 cc. of strong sulphuric acid. A 
small quantity of mercury or mercury salt is added, together with a small piece 
of paraffine wax. The mixture is heated until che rubber is dissolved and the 
black liquid begins to clear. Two to 4 grams of potassium sulphate are then 
added and the heating continued until a colorless or pale yellow liquid is obtained. 
After cooling, 1 to 2 grams of potassium metabisulphite are added and an excess 
of tartaric acid. The liquid is diluted sufficiently to prevent the charring of 
the tartaric acid and boiled until the odor of sulphurous acid has disappeared. 
A few cc. of dilute hydrochloric acid are added, the liquid diluted to 200 cc, 
filtered through a dry filter, and 195 cc. titrated either with iodine or with 
potassium bromate (the latter in acid solution), as described under the volu- 
metric procedures. 

SEPARATIONS 

Separation of Antimony (together with Members of the Hydrogen Sul- 
phide Group), from Iron, Chromium, Aluminum, Cobalt, Nickel, Manganese, 
Zinc, the Alkaline Earths, and Alkalies. The acid solution of the elements 
is saturated with hydrogen sulphide, the elements of the Hydrogen Sulphide 
Group are precipitated as sulphides, the other elements remaining in solution. 
Antimony sulphide may be precipitated from an hydrochloric acid solution con- 
taining 15 cc. of strong acid per 100 cc. of solution; lead and cadmium are 
incompletely precipitated. 

Separation of Antimony (together with Arsenic and Tin), from Mer- 
cury, Copper, Bismuth, Cadmium and Lead. The sulphides of antimony, 
arsenic, and tin are soluble in a mixture of sodium hydroxide and sodium sul- 
phide, the soluble sulpho salts being formed, mercury, copper, bismuth, cadmium, 
and lead remaining as insoluble sulphides. The following procedure may be 
used for alloys free from members of other groups. The acid solution is treated 
with 3 to 5 grams of tartaric acid and diluted slightly (more tartaric acid being 
added if the solution becomes turbid), then poured into 300 cc. of a mixture of 
sodium sulphide and sodium hydroxide (150 cc. of the mix described under 

1 W. Schmitz, Chem. Zentralbl., 1911, ii, 1710. Analyst, 1912, p. 64. 



22 ANTIMONY 

"Solution of Sulphide Ores" diluted to 300 cc). The mixture is wanned and 
the insoluble sulphides allowed to settle out. The solution is filtered free of 
the precipitate and the latter washed. The filtrate is acidified with hydro- 
chloric or sulphuric acid and saturated with hydrogen sulphide. The sul- 
phides of arsenic, antimony and tin are now filtered off and treated as described 
later. 

Separation of Arsenic, Antimony, and Tin. The sulphides may be dis- 
solved in concentrated hydrochloric acid by addition of potassium chlorate to 
oxidize the sulphur to sulphuric acid. This oxidation may be effected in the 
alkaline solution of the sulpho salts by addition of 30% hydrogen peroxide 
in small portions until the yellow solution is completely decolorized and then 
1 to 2 cc. in excess, the solution then boiled to completely oxidize the sul- 
phides to sulphates and to remove the excess of peroxide. The solution is 
then acidified, the precipitation of the sulphides and the subsequent filtration 
and resolution being avoided. 

Removal of Arsenic. This may be accomplished by volatiUzing arsenic as 
arsenic trichloride in a strong hydrochloric solution by boiling. If arsenic is 
to be determined the procedure given under the chapter on arsenic is followed, 
the arsenic being distilled in a current of hydrochloric acid gas. If arsenic 
is not desired it may be expelled by reducing the solution with sodium meta- 
bisulphite or potassium iodide and boiling. Antimony and tin remain in the 
concentrated acid solution. 

The separation of arsenic from antimony and tin may be effected by remqval 
of the former in a strong hydrochloric acid solution as described imder the section 
"Preparation and Solution of the Sample," arsenic being precipitated by hydrogen 
sulphide, whereas antimony and tin remain in solution. 

Separation of Antimony from Tin. Upon the removal of arsenic, anti- 
mony may be determined directly in the presence of tin by one of the volu- 
metric methods given later. If a gravimetric separation is desired, it may 
be made according to a modification of Clark's method, * which depends upon the 
fact that antimony is completely precipitated from a solution containing oxalic 
acid, by hydrogen sulphide, whereas tin is not. The tin must be in the stannic 
form, otherwise the insoluble crystalline stannous oxalate will form. 

If the mixture is acid, it is neutralized with caustic and twenty times the 
weight of the Sn and Sb present added in excess, e.g., 2 grams potassium 
hydroxide in excess for every 0.1 gram of tin and antimony present in the solu- 
tion. About ten times as much of tartaric acid is now added as the maximum 
weight of the two metals, followed by 30% hydrogen peroxide to oxidize the 
tin. The excess of peroxide is removed by boiling. To the slightly cooled 
solution a hot solution of pure oxalic acid is added, 5 grams of oxalic acid for 
each 0.1 gram of the mixed elements. COj+Oj are evolved. The solution 
is boiled for about ten minutes and the volume made up to about 100 cc. 
Hydrogen sulphide is rapidly passed into the boiling solution until a change 
from a white turbidity to an orange color takes place and antimony begins 
to precipitate. The passage of the gas is continued for fifteen minutes, the 
solution diluted with hot water to a volume of 250 cc. and hydrogen sulphide 
passed into the boiling solution for another fifteen minutes. The flame is now 
removed and the HjS " gasing " continued for ten minutes longer. The pre- 
cipitated antimony pentasulphide is filtered off in a weighed Gooch crucible. 
' The Original procedure may be found in Chem. News, Vol. XXI, p. 124. 



ANTIMONY 23 

It may be determined gravimetrically as SbaSj, according to the procedure 
given later, by washing with 1% oxalic acid and dilute acetic acid, by decan- 
tation, the solutions being hot and saturated with hydrogen sulphide. The 
precipitate washed into the crucible is dried in a current of COa at a heat of 
280 to 300° and weighed as ShS». 

Tin may be determined electrolytically in the filtrate evaporated to about 
150 cc, the oxalic acid being nearly neutralized with ammonia. See Electro- 
lytic Determination of Tin. 

Antimony may be separated from tin in a hot hydrochloric acid solution 
by addition of pure iron. The iron and tin sulphides are dissolved in concentrated 
hydrochloric acid plus a few crystals of potassium chlorate. The solution 
should contain about 10% hydrochloric acid, more hydrochloric acid being added 
as the iron dissolves. Antimony is precipitated as a metal. 



GRAVIMETRIC METHODS FOR THE DETERMINATION OF 

ANTIMONY 

The accuracy and rapidity of volumetric methods for the determination 
of antimony leave little to be desired in the estimation of this element, so tliat 
the more tedious gravimetric methods are less frequently used. The following 
procedures are given in view of possible utility in certain analyses. 

Determination of Antimony as the Trisulphide, Sb2S3 ^ 

Although hydrogen sulphide passed into a cold solution tends to precip- 
itate SbjSs, in hot strongly acid solutions, the lower sulphide, SbaSs, tends to 
form. The higher sulphide is decomposed at 230° C. with formation of Sb2S3 
and the volatilization of sulphur. A temperature* of 280 to 300° is even more 
favorable for this transformation. The method takes advantage of these con- 
ditions for formation of antimony trisulphide, in which form it is weighed. 

Procedure. The solution of antimony, free from arsenic, is treated in an 
Erlenmeyer flask with strong hydrochloric acid until the solution contains about 
20% of the concentrated acid. The mixture is heated to boiling and a slow 
current of hydrogen sulphide is passed into the hot solution until the precipitate 
passes from a yellow color through an orange and finally becomes a dark red 
to black color. The flask is agitated gently to coagulate the precipitate, which 
settles in a crystalline form. The solution is diluted with an equal volume of 
water, washing down the walls of the flask. A slight turbidity is generally 
seen, due to precipitation of a small amount of antimony that remains in solu- 
tion in a strong acid solution. H2S is now passed into the diluted solution 
until it becomes clear, thirty-five to forty minutes are usually sufficient to pre- 
cipitate all of the antimony. The precipitate is transferred to a weighed Gooch 
crucible, washed with small portions of water containing hydrogen sulphide, 
and finally with pure water. 

It is a common practice, at this juncture, to wash the precipitate with car- 
bon disulphide or carbon tetrachloride to remove precipitated sulphur. Alcohol 
is now used, followed by ether, and the precipitate sucked dry. 

* Method of Vortmann and Metzel modified. 
« Paul, Z. anal. Chem. SI, 540 (1892). 



24 ANTIMONY 

The Gooch crucible is placed in a large combustion tube and heated in a 
current of dry, pure COi at 130° C. for an hour. The temperature is now 
raised to 280 to SOO'' C. and the heating continued for two hours. The residue 
will consist of pure SbjSj. 

Sb2S,X0.7l42 = Sb, or Sb^SaX 0.8568 =Sb,0,. 

Notes. Antimony may be determined by oxidation of the sulphide precipitate 
by means of fumine nitric acid. The mixture evaporated to dryness is ignited and 
the residue weighed as SbjO*. The temperatiu*e of the ignition should be between 
750 to 800** C. The volatile trioxide forms at a little above 950°. The procedure 
requires greater care than the sulphide method and possesses no advantages. 

Pure carbon dioxide may be obtained from limestone placed in a Kipp generator. 
The gas is dried by passing it through strong sulphuric acid. It shoulcl oe free from 
oxygen of the air. It is advisable to sweep out the air from the generator before 
attaching it to the combustion train. The air in the tube is swept out with carbon 
dioxide before heating the sample. 

Property of SbiSj, m.w.f 336.61; 8p.gr. ^ 4.65; fusible and volatile; solubility, 
0.000175 gram per 100 cc. HaO; decomposed by hot HjO; soluble in alkalies, NH4HS, 
K,S, cone. HCl. 

Electrolytic Determination of Antimony ^ 

The chief condition for the success of the electrolytic deposition of antimony 
in metallic form is the absence of polysulphides, since these substances prevent 
the element from being deposited, 2Sb+3NaxS2=2Na8SbSj. The formation 
of polysulphides may be prevented during electrolysis by addition of potassium 
cyanide to the solution,* NajSj+KCN =Na4S+KCNS. 

The results of this method, according to F. Henz,» are invariably 1.5 to 
2% too high of the total antimony present in the solution. Treadwell and 
Hall recommend subtraction of a constant factor of 1.6% of the weight of the 
antimony deposited. The sample for analysis should contain not over 0.2 gram 
antimony. 

Procedure. Antimony precipitated as the sulphide is washed and then 
dissolved off the filter by pouring pure sodium sulphide solution (sp.gr. 1.14) 
over the precipitate, the solution being caught in a weighed platinum dish, 
with unpolished inner surface. The total volume of the solution should be 
not over 80 cc. (if less than this, additional NajS solution is added to make 
up to 80 cc). Sixty cc. of water followed by 2 to 3 grams of potassium cyanide 
(C.P.) are added and the cyanide dissolved by stirring with the rotating anode. 
The solution heated to 60 to 70° is electrolyzed with a current of 1 to 1.5 amperes, 
E.M.F. =2 to 3 volts. Two hours are generally sufficient to deposit all the 
antimony. The light-gray deposit adheres firmly upon the cathode. With- 
out breaking the current the solution is siphoned off, while fresh water is 
being added, until the current ceases to flow through the liquid. The cathode 
is washed thoroughly with water, followed by alcohol and ether and then dried 
at about 80°, cooled in a desiccator and weighed. 

The antimony deposits may be removed by heating with a solution of alkali 

* Method -first proposed by Parrodi and Mascazzini, Z. anal. Chem., 18, 587 
(1879), modified by Luckow, Z. anal. Chem., 19, 13 (1880), and later improved 
by Classen and Reiss, Berichte, 14 1629 (1881); 17, 2474 (1884); 18, 408 (1885); 
27, 2074 (1894). * Treadwell and Hall, Analytical Chemistry. «Z. anorg. Chem., 
87, 31 (1903). 



ANTIMONY 25 

polysulphide or by a mixture of equal parts of saturated solution of tartaric 
acid and nitric acid. 

VOLUMETRIC METHODS 

Potassium Bromate Method for Determining Antimony ^ 

Outline. This method is of special value in determining antimony in hard 
lead and alloys.* It was first suggested by Gyory and later modified by Siedler, 
Nissensen and Rowell.' The process is based upon the oxidation of antimony 
from the trivalent to the pentavalent form by potassium bromate, the follow- 
ing reaction taking place: 

KBr03+3SbCU+6HCl =3SbCU+KBr+3H,0. 

Standard Solutions. 

Antimony Chloride Solution, Six grams of the C. P. pulverized metal are 
dissolved in 500 cc. of concentrated hydrochloric acid together with 100 cc. 
saturated bromine solution, more acid and bromine added if necessary to effect 
solution. After expelling the bromine by boiling, about 200 cc. concentrated 
hydrochloric acid are added and the whole made up to one liter. Fifty cc. 
=0.3 gram antimony. 

N/10 Potassium Bromate Solution, 2.82 grams of C. P. salt are dissolved 
in water and made up to 1 liter. Theoretically 2.7852 grams are required, but 
the salt invariably contains potassium bromide as an impurity. The solution 
is standardized against 50 cc. of the antimony chloride solution, which has 
been reduced with sodium sulphite according to the standard scheme. One 
cc. of N/10 KBrO, =0.006 gram Sb. 

Methyl Oranye, 0.1 gram M. O. per 100 cc. of distilled water. The indi- 
cator should be free from sediment. 

Saturated Bromine Solution, 500 cc. concentrated hydrochloric acid satu- 
rated with 70 cc. of bromine. 

Procedure. Solution, One gram of the finely divided alloy is brushed 
into a 500-cc. beaker, 100 cc. of concentrated hydrochloric acid and 20 cc. of 
saturated bromine solution are added. The beaker is covered and placed on 
the steam bath xmtil the metal dissolves. It may be necessary to add more 
bromine and acid to effect complete solution. In case the oxides of antimony 
and tin separate out and do not redissolve, fusion with sodium hydroxide may 
be necessary. Bromine is now expelled by boiling the solution down to about 
40 cc. 

Reduction, One hundred cc. of concentrated hydrochloric acid and 10 cc. 
of a fresh saturated solution of NajSOs are added and the solution boiled down 
to 40 cc, on a sand bath, to expel arsenic and the excess of normal sodium sul- 
phite. Samples high in arsenic may require a repetition of the reduction. 

Titration, The cover and sides of the beaker are rinsed down with 20 cc. 
of hydrochloric acid (sp.gr. 1.2) followed by a few cc. of hot water and the 
solution heated to boiling on a sand bath. The standard bromate solution 
is now run into the hot solution of antimony to within 2 to 3 cc. of the end- 

»S. Gydry, Zeit. Anal. Chem., 82, 415 (1893). J. B. Dimcan, Chein. News, 95, 
49 (1907). 

* H. W. Rowell, Jour. See. Chem. Ind., XXV, 1181. 



26 ANTIMONY 

point, this having been determined in a preliminary run with methyl orange 
added in the beginning, 4 drops of methyl orange are added and the titration 
completed cautiously until the color of the indicator is destroyed. If iron 
or copper is present the final product will appear yellow. Since the end-reaction 
is slow the last portion of the reagent should be added drop by drop with con- 
stant stirring. 

1 cc. N/10 KBrOs =0.006 gram Sb. 

Notes. Since antimony chloride begins to volatilize at 195° C. and boils at 220° C. 
it is advisable not to carry the concentration too far while expelling arsenic. 

Lead, copper, zinc, tin, silver, chromium, and sulphuric acid nave no effect upon 
the determination, but large quantities of calcium, magnesium, and anunonmm 
salts tend to make the results high. Low^ found that copper proauced high results, 
approximately .012% too high for every 0.1% of copper present. The author (W.W.S.) 
finds, however, that with the procedure given above, amounts of copper as high as 
15% producea no diflSculty beyond a j'^ellow coloration of the solution. With larger 
amounts of copper, the end-point became difficult to detect owing to the depth of 
this yellow color, so that in case of brass and copper alloys, the method must be 
modified by a procedure for removal of the copper. I^ead up to 95% caused no 
difficulty. Iron, in amounts such as are commonly met in alloys of lead, does not 
interfere. 

During the course of analysis antimony may be isolated as the sulphide* this is 
dissolved m strong hydrochloric acid, and reduced and concentrated to expel arsenic 
that may be present as a contamination, and the resulting solution titrated with 
potassium bromate as directed above. 

Sources of Error, (a) Imperfect volatihzation of arsenic. (6) Incomplete expul- 
sion of SOj. (c) Over-titration if insufficient hydrochloric acid is present. 

No loss of antimony occurs at temperatures below 120° C. 

Potassium Iodide Method for Determining Antimony 

Protedure. To 1 gram of fine sawings or filings in a 16-oz. Erlenmeyer flask 
add 60 cc. of concentrated hydrochloric acid and heat on an asbestos board or on 
the water bath just below boiling. When hydrogen is no longer evolved, decant 
the liquor and wash twice with concentrated hydrochloric acid, retaining the 
antimony in the flask. Now dissolve the antimony by adding 15 cc. of con- 
centrated hydrochloric acid and solid potassium chlorate, a few crystals at a time, 
until the antimony is in solution, the liquid being kept hot. Expel chlorine 
by boiling, add 60 cc. of concentrated hydrochloric acid and again bring 
to boiling. Cool and add 20 cc. of 20% potassium iodide solution and 1 cc. of 
carbon disulphide or tetrachloride. Titrate the liberated iodine with tenth-normal 
sodium thiosulphate. The brown color will gradually disappear from the solu- 
tion and the last traces of free iodine will be collected in carbon disulphide or 
carbon tetrachloride, giving a pink color. When this pink color disappears the 
end-point has been reached. 

One cc. N/10 Na^2Oi=^.006 gram of Sb. 

NasSsOs is standardized against .3 gram antimony as in case of Potassium 
Bromate Method, the above procedure, however, being followed. Antimony 
must be free from copper and arsenic. 

Notes. The following reversible reaction is of interest: "R" representing a tri- 
valent metal with oxidation to pentavalent form. 

Rj08+2l2+2H,0 =RtOfc+4HI. 

The reaction goes to the right when an alkali is present to neutralize the free 
acid formed; I e.g., Mohr's process for determining arsenic by titration of the lower 

* A. H. Low, "Technical Methods of Ore Analysis." 



ANTIMONY 27 

oxide with iodine in presence of sodium bicarbonate. The reaction goes to the left 
in presence of strong acid; e.g., Weller's process for the determination of antimony 
in an acid solution. 

The solution should not contain more than i of its volume of hydrochloric acid 
(sp.gr. 1.16), since too much hydrochloric acid gives high results, owing to the action 
of hydrochloric acid on potassium iodide. Too little acid leads to the separation 
of basic iodides and chlorides of antimony. The solution is best boiled down to 
20% hydrochloric acid (above strength). 

Stannous chloride may be used in place of thio-sulphate in titration of iodine. 

SbCU+2KI=SbCl5+2KCl+I, and I,+SnCl,+2HCl=SnCl4+2HI. 

Determination of Antimony by Oxidation with Iodine 

The procedure originated by Mohr and modified by Clark, depends upon 
the reaction Sb,0,-|-2I,-|-2H,0 =Sb206+4HI. 

The reaction takes place when iodine is added to a solution of antimonous 
salt in presence of an excess of alkali bicarbonate. In an acid solution oxida- 
tion with iodine does not go beyond SbiOs. 

Procedure. Solution. The sample is brought into solution by one of the 
procedures given under "Preparation and Solution of the Sample.^' Alloys 
of antimony, lead, and tin are treated according to directions given for this 
combination. 

Titration. To the hydrochloric acid solution of antimony is added tar- 
taric acid or Rochelle salt^, the excess of the acid neutralized with sodium car- 
bonate, the solution made barely acid with hydrochloric acid and a saturated 
solution of sodium bicarbonate added in the proportion of 10 cc. bicarbonate 
solution for each 0.1 gram of SbiOj. Starch is added as an indicator and the 
solution titrated with N/10 iodine. 

1 cc. N/10 iodine =0.006 gram Sb. 

Note. The titration should be made inmiediately upon addition of the sodium 
salts. 

Antimony in Solder Metal and Alloys with Tin and Lead ^ 

Procedure. Dissolve 2 grams of the sample of alloy in concentrated hy- 
drochloric acid. When the metal is all in solution, add crystals of iodine until 
the solution is thoroughly permeated. The color at this point should be a 
deep purple. Boil until all of the iodine fumes have been driven out. The 
metallic antimony which did not go into solution in the hydrochloric acid should 
now be all dissolved. If it is not, add more iodine until the solution is com- 
plete. When all is in solution and the color changes to a straw yellow, cool, add 
a few cc. of starch solution. If a blue color appears, due to an excess of iodine, 
run in N/10 sodium thiosulphate solution until colorless. In case there is no 
blue color developed, add N/10 iodine until a faint blue appears. Now add 
50 cc. of a saturated solution of Rochelle salts. Make alkaline to litmus by 
adding 25% sodium hydrate solution. Then make slightly acid with HCl 
and finally alkaline with sodium bicarbonate. Cool and titrate with N/10 
iodine. 

Note. " The method gives very ^ood results. I have checked it up when there 
was one-tenth of a gram known antimony present and the results were within a 
reasonable limit of acciu-acy." ^ 

1 Method conmiunicated to author by Mr. B. S. Clark. 



28 ANTIMONY 

Other Procedures 
Permanganate Method 

Antimonous salts may be titrated with standard potassium pennanganate. 
The iron value for the permanganate multiplied by 1.075 or the oxalic acid 
(C2H2O4 -21120) value multiphed by 0.9532, will give the antimony value.^ 

Indirect Evolution Method 

The method depends upon the evolution of H2S from the sulphides of anti- 
mony decomposed by strong hydrochloric acid, the amount of hydrogen sul- 
phide being the same for either Sb2S3 or Sb2S6, the following reactions taking 
place: 

1. Sb2S,+6HCU2SbCl,+3H2S. 

2. Sb2S5+6HCU2SbCh+S2-f3H2S. 

The details of the method are practically the same as determination of 
sulphur by the evolution method in the analysis of iron and steel. See Chapter 
on Sulphur. The antimony sulphide precipitate is placed in the evolution 
flask, strong hydrochloric acid added with an equal volume of water and the 
evolved hydrogen sulphide absorbed in an ammoniacal solution of cadmium 
chloride. The precipitated cadmium sulphide is then titrated with iodine in 
an acid solution. 

One cc. N/10 1=0.001604 gram S, since 3S=2Sb, therefore Sb=SX2.499, 
hence, 1 cc. N/10 I =0.00401 gram Sb. 

Preparation of Standard Iodine Solution. An approximate tenth normal solu- 
tion is made by dissolving 12.7 grams of commercial iodine, roughly weighed on 
a watch-glass, in 200 cc. of water containing about 25 grams of potassium iodide, 
solution being effected in a graduated liter flask. After making up to 1000 cc. 
with distilled water, the reagent is transferred to a dark-colored bottle, to protect 
it from light. It is advisable to make up 5 to 10 liters at a time for laboratories 
where the solution is in constant demand. After standing several hours, the 
reagent is standardized by running a portion from a burette into 100 cc. of tenth 
normal arsenous acid (see page 204) until a faint yellow color is perceptible. In 
presence of starch indicator a faint blue color is obtained. 

100 divided by the cc. of iodine required gives the factor for aN/10 solution. 

Example. If 98.5 cc. of iodine are required, 100-^98.5 = 1.0152 N/10 or 
.10152 normal. 

Tenth normal arsenous acid solution contains 4.953 grams of A820a, per liter, dis- 
solved in sodium hydroxide and made up according to directions ^ven on page 204. 
The oxide is seldom pure, so that allowance must \ye made for impurities. For example, 
the acid containing 99.56 per cent AS2O3 would require 4.953 -r .9956 =4.97 grams per 
liter of solution. 

Commercial iodine may contain chlorine, bromine, cyanogen and water. It may l)e 
purified by repeated sublimation ('* Analytical Chemistry," TreadwcH and Hall, IV 
Ed., page 646, or "A Treatise on Quantitative Inorganic Analysis" (1913), by J. W. 
Mellor, page 288). lliere is no advantage in taking the theoretical amount of purified 
iodine, however, since the reagent changes in strength on standing. 

Potassium iodide augment^s Rolution of iodine, which is sparingly soluble in water. 

The iodine may l>e standardized by titrating a definite volume with N/10 sodium 
thiosulphate. See page 204. 

* Technical Methods of Ore Analysis, by A. H. Low. 



ANTIMONY 29 

Determination of Small Amounts of Antimony^ 

The determination depends upon the fact that when antimony compounds 
are acted upon in acid solution by nascent hydrogen the gas stibine is evolved, 
which forms a black compound with silver nitrate. The method is very similar 
to Allen and Palmer's modification of the Gutzeit procedure for arsenic, dif- 
fering, however, in the facts that heating is necessary to evolve completely 
the stibine, the presence of iron is not required, and stannous chloride is not 
used. 

Procedure. The material is brought into solution with water, or by treat- 
ment with hydrochloric acid, or hydrochloric acid and an oxidizing agent (KClOj or 
Br) with subsequent evaporation to dryness on the steam plate or water bath, or 
by fusion with sodium carbonate followed by acid extraction. If arsenic is present, 
the solution, contained in a distillation flask, is reduced with ferrous sulphate or 
chloride and arsenic distilled off in a current of HCl gas, according to the proced- 
ure outlined under Arsenic. The volume of the solution is reduced from 
about 200 cc. to 50-€0 cc* Antimony is now isolated by continuing the distilla- 
tion with addition of zinc chloride to raise the boiling-point of the solution. 
Thirty cc. of a saturated solution of zinc chloride are added to the Uquor in the 
flask and antimony distilled, hydrochloric acid (sp.gr. 1.2) being added through 
a separatory funnel, drop by drop, to replace the solution evaporated. The first 
fifty cc. of the distillate will contain all the antimony, present in small amount. 
The excess of acid is carefully neutraUzed with sodium carbonate, leaving the 
solution slightly acid. The mixture is placed in the modified Gutzeit apparatus, 
pure zinc shot added, and the apparatus connected up as described for deter- 
mining traces of arsenic. Fig. 2. In place of the mercuric chloride, silver 
nitrate paper is used for obtaining the stain, as this reagent is more sensi- 
tive to stibine. The apparatus is placed in warm water (about 60** C.) for 
two hours. The silver nitrate paper is then removed, inmiersed in 10% solu- 
tion of sodium thiosulphate to fix the stain and washed with distilled water 
to remove the silver nitrate. The paper is then dried and compared with a 
standard set of stains made by placing known amounts of antimony in solu- 
tions of like material examined, and proceeding according to the outline given. 

Notes. Potassium antimony! tartrate may be taken for the standard antimonv 
solution. 0.2765 gram of the salt is diluted to a liter and 10 cc. of this stock solution is 
diluted to 1 liter. 10 cc. of this final solution is equivabnt to 0.01 milligram of anti- 
mony. Stains representing 0.005 to 0.05 milligram antimony are suited for the test. 

Silver nitrate paper. This is made bv dipping sheets of filter paper in a 0.4% 
solution of silver nitrate, running through the mangle to remove the excess of silver 
nitrate, drying and cutting into strips according to the procedure recommended for 
the paper useof in the Gutzeit method for arsenic. 

Blank runs should be made with the reagents and the blanks deducted from the 
stains obtained in the regular tests. If possible, arsenic and antimony-free reagents 
should be used. 

The author is indebted to Mr. J. P. Kelly for his review and criticism of this chapter. 

* Method suggested bv C. R. Sanger, communicated to the author by Mr. J. P. Kelly. 

* During the removal of arsenic the temperature of the solution should not rise 
above 125° C., since a loss of antimony occurs al)ove this point. It is advisable^ there- 
fore^ to place a thermometer in the flask, and observe the temperature durmg the 
distillation. 



ARSENIC 

Wilfred W. Scott 

AB,at.wt. 74.96- ^'•«'«'- 8p.gr. ^'^^ m.p. ^^ b.p. «"^'- ^' 

amorp. 4.72 .... < 360° 

Oxides, AssOs, As^Of^ 

DETECTION 

Hydrogen sulphide precipitates the yellow sulphide of arsenic, AssSs, when 
passed into its solution made strongly acid with hydrochloric acid. If the 
solution contains more than 25% hydrochloric acid, (sp.gr. 1.126) the other 
members of the hydrogen sulphide group do not interfere, as they are not 
precipitated from strong acid solutions by hydrogen sulphide. Arsenic sulphide 
is soluble in alkaline carbonates. (Antimony sulphide, Sb^Sa, reddish yellow, 
is insoluble in alkaline carbonates.) 

Volatility of the chloride, AsClt, is a means of separation and distinction 
of arsenic. Details of the procedure are given under "Separations." The 
distillate may be tested for arsenic as directed above. 

Traces of arsenic may be detected by either the Gutzeit or Marsh test 
for arsenic. Directions for the Gutzeit test are given at the close of the vol- 
umetric procedures. 

Distinction between Arsenates and Arsenites. Magnesia mixture pre- 
cipitates white, MgNHiAsOi, when added to ammoniacal solutions containing 
arsenates, but it produces no precipitate with arsenites. 

Red silver arsenate and yellow silver arsenite are precipitated from neutral 
solutions by iimmoniacal silver nitrate. An arsenate gives a yellow precipitate 
with ammonium molybdate solution. 

ESTIMATION 

The determination of arsenic is required in the valuation of native arsenic, 
white arsenic, As.O*; ores of arsenic — orpiment, AsiSi; realgar, AS2S1; pyrar- 
gjrrite, AssSbs; arsenopyrite, or mispickel, FeSAs; cobaltite or cobalt glance, 
CoSAs; smaltite, CoAsi; niccoHte, NiAs. The substance is estimated in 
copper ores, in speiss, regulus; in iron precipitates (basic arsenate). It is 
determined in paint pigments, ScheeFs green, etc. The element is determined 
in shot alloy and in many metals. It is estimated in germicides, disinfectants, 
and insecticides — Paris green, lead arsenate, zinc arsenite. Traces are looked 
for in food products and in substances where its presence is not desired. 

Preparation and Solution of the Sample 

In dissolving arsenic compounds it will be recalled that the oxide, As^Oj, 
is not readily acted upon by dilute acids — hydrochloric or sulphuric. The 
compound is soluble, however, in alkaline hydroxides and carbonates. Nitric 

^Van Nostrand's Chem. Annual — Olsen — 3d Ed. 

30 



ARSENIC 31 

acid oxidizes AstOa to the higher oxide, AssOi, which is soluble in water. The 
sulphides AstSs and As& are practically insoluble in hydrochloric or sulphuric 
acids, but are dissolved by the fixed alkalies and alkali sulphides. All arsenites, 
with the exception of the alkali arsenites, require acids to effect solution. 

Pyrites Ore and Arseno-pyrites. The amount of the sample may vary 
from 1 to 20 grams,* according to the arsenic content. The finely ground sample 
in a large casserole is oxidized by adding 10 to 50 cc. of bromine solution (75 
cc. KBr+50 cc. liquid Br+450 cc. H2O) covering and allowing to stand for 
fifteen minutes, then 20 to 50 cc. of strong nitric acid are added in three or 
four portions, allowing the action to subside upon each addition. The glass 
cover is raised by means of riders, and the sample evaporated to dryness on 
the steam bath; 10 to 25 cc. of hydrochloric acid are now added and the sample 
again taken to dryness. Again 10 to 25 cc. of hydrochloric acid are added 
and the sample tiJcen to dryness. Finally 25 cc. of hydrochloric acid and 75 
cc. of water are added, and the mixture digested over a low flame until all the 
gangue, except the silica, is dissolved. The solution is now examined for arsenic 
by distillation of the arsenic after reduction, the distillate being titrated with 
standard iodine solution according to directions given later. 

Arsenous Oxide. The sample may be dissolved in caustic soda, the solution 
neutralized with hydrochloric acid, and the resulting sample titrated with iodine. 

Arsenic Acid, Alkali Arsenates, etc. The sample is dissolved in 20 to 25 
cc. of dilute sulphuric acid, 1 : 1, in an Erlenmeyer flask, and reduced by 
addition of 3 to 5 grams of potassium iodide, the action being hastened by placing 
the mixture on a steam bath. The iodine liberated is exactly neutralized with 
thiosulphate and the arsenous acid titrated with iodine according to the pro- 
cedure given later. If a N/10 iodine solution is to be used, the sample should 
not contain over .37 gram arsenic. A 10-gram sample may be taken, made 
up to a definite volume and aliquot parts taken for analysis. 

Arsenic in Sulphuric Acid. Arsenous acid may be titrated directly with 
iodine in a 20- to 50-gram sample, which has been diluted to 200 to 300 cc. 
with water and nearly neutralized with ammonium hydroxide and then an excess 
of sodium acid carbonate added, followed by the iodine titration.* 

Arsenic Acid in Sulphuric Acid. Twenty-five cc. of the acid containing 
about 0.1% arsenic or a larger volume in case the percentage of arsenic is less 
than 0.1% AsjOi (the sp.gr. of the acid being known) are measured out into a 
short-necked Kjeldahl flask. About half a gram of tartaric acid and 2 grams of 
fused, arsenic-free potassium bisulphate are added and the acid heated over 
a low flame until the liberated carbon is completely oxidized and the acid again 
becomes clear, e.g., a pale straw color. It is not advisable to heat to violent 
fuming, as a loss of arsenic is then apt to occur. The cooled acid is poured 
into about 300 cc. of water, the excess acid nearly neutralized with ammonia, 
bicarbonate of soda added in excess and the arsenous acid titrated with standard 
iodine. Total arsenic as As20t minus arsenous arsenic as As203== arsenic arsenic 
in terms of As^Os. This result multiplied by 1.1616 ^AsjOj. 

Arsenic in Hydrochloric Acid. The arsenic in 20 to 100 cc. sample is 
reduced by ferrous chloride, the arsenic distilled according to directions given 
later, and the distillate titrated with iodine. 

* 0.1% arsenic determined on a 20-gram sample. 

^SOj should be expelled by heat or by a current of air before treating with the 
alkali. 



32 ARSENIC 

Arsenic in Organic Matter. ^ 0.2 to 0.5 gram of the sample finely powdered 
is oxidized by mixing with 10 to 15 grams of sodium carbonate and sodium 
peroxide, 1 : 1, in a nickel crucible, a portion of the fusion mixture being spread 
over the charge. After heating gently for fifteen minutes, the fusion is com- 
pleted by heating to dull redness for five minutes longer. The contents of 
the crucible are rinsed into an Erlenmeyer flask after extraction with water, 
and the solution made acid with dilute sulphuric acid, 1:1. The mixture 
is boiled down to 100 cc, 1 to 2 grams of potassium iodide added and the solu- 
tion further concentrated to about 40 cc. Iodine is reduced with sulphurous 
acid or thiosulphate, the solution diluted with hot water and saturated with 
hydrogen sulphide. Arsenous sulphide is filtered off, washed, dissolved in 15 
to 20 cc. of half-normal sodium hydroxide and 30 cc. of hydrogen peroxide (30%) 
solution added, and the solution boiled. About 12 cc. of dilute sulphuric acid, 
1:1, are added, together with 1 to 2 grams of potassium iodide, the solution 
concentrated to 40 cc. and free iodione reduced with thiosulphate as before. 
Arsenic is now titrated, with standard iodine, upon neutralization of the free acid 
with sodium hydroxide and sodium acid carbonate. 

Lead Arsenate. Ten grams of the thoroughly mixed paste or 5 grams of 
the powder are dissolved by treating with 25 cc. of 10% hot sodium hydrox- 
ide solution, and diluted to 250 cc. An aliciuot part, 50 cc. (=2 grams paste 
and 1 gram powder) is placed in an Erlenmeyer flask and 20 cc. of dilute 
sulphuric acid, 1:1, added, and the solution diluted to 150 cc. About 3 grams 
of solid potassium iodide are added and the solution boiled down to about 
50 cc. (but not to fumes). The liquor will be colored yellow by free iodine. 
Tenth normal sodium thiosul])hate is added drop by drop until the free iodine 
is neutralized (solution loses its yellow color), it is now diluted to about 250 
cc. and the free acid neutralized by anmionium hydroxide (methyl orange 
indicator), then made slightly acid with dilute sulphuric acid, and an excess 
of bicarbonate of soda added. The arsenic is titrated with standard iodine. 

The arsenic may be rt^duccd by placing the 50-cc. sample in a Kjeldahl 
flask, adding 25 cc. of strong sulphuric acid (1.84 sp.gr.), | gram tartaric acid 
and 2 grams acid potassium sulphate, KHSO4, and digesting over a strong flame 
until the organic matter is destroyed and the solution is a pale yellow color. 
The cooled acid is diluted and neutralized, etc., as directed above. 

Water-soluble Arsenic in Insecticides. Rapid Works Test. Two granis 
of the paste is digested with 1000 cc. of water at 90** C. for five minutes, in 
a graduated 1000-cc. flask. An ali<iuot portion is filtered and the arsenic 
detennined by the (Jutzeit method. 

Water-soluble arsenite may be titrated directly with iodine in presence 
of sodium bicarbonate. 

Zinc Arsenite. About 5 grams of powder or 10 grams of paste are taken 
and dissolved in a warm solution containing 300 cc. of water and 25 cc. of 
strong hydrochloric acid. The cooled solution is diluted to 500 cc. and 100-cc. 
portions taken for analysis. The acid is partly neutralized with ammonium 
hydroxide and 50 cc. of a saturated solution of ammonium oxalate added 
(to prevent i)recipitati()n of the zin<" as ZnCOs), and an excess of sodium 
bicarl)onate, NaHCOs. Arsenic is now titrated with iodine as directed later. 

Soluble Arsenic in Zinc Arsenite. One gram sam])le is rubbed into an 

* Little, Cahan, and Morgan, Jour. Chem. Soc, 96, 1477 (1909). 



ARSENIC 33 

emulsion with several portions of water until the whole is in suspension. The 
cloudy liquor is diluted to 1000 cc. and a portion filtered through a }-in. 
asbestos mat on a perforated plate, the asbestos being covered with a layer 
of filter paper. The first 50 cc. are rejected. One hundred cc. of the clear 
filtrate (=0.1 gram) is treated with 10 cc. of strong sulphuric acid, 0.05 gram, 
Fe^Os (use ferric ammonium sulphate) and i cc. of 80% stannous chloride solution 
and heated until colorless. Arsenic is now detennined by the Gutzeit method, 
using the larger-sized apparatus. 

Arsenic in Mispickel. One gram of the finely powdered mineral is fused 
in a nickel crucible with about 10 grams of a mixture of potavssium carbonate 
and nitric acid, 1 : 1,* and the melt extracted with hot water. Two hundred cc. 
of a saturated solution of SO2 is added to the filtrate to reduce the arsenic, 
the excess of SOj then expelled by boilinp, the solution diluted with dilute 
sulphuric acid, and arsenic determined in the filtrate. 

Arsenic in Steel, Iron, Pig Iron, etc. One to 50 grams of steel, etc., may 
be treated according to the scheme for pyrites. If a large sample is taken, 
it is advisable to treat it in a 500-cc. flask, connected with a second flask 
containing bromine, to guard against loss of arsenic by volatilization. When 
the sample has dissolved it is taken to dryness (the bronune in the second flask 
being combined with it) and treated as directed in i)yrites. Arsenic chloride, 
AsCU, is transferred to the dlstillnig flask with strong hydrochloric acid, and 
arsenic separated from the iron by volatilization of reduced chloride accord- 
ing to the procedure given below. 

Arsenic in Copper. Arsenic Ls precipitated with iron by the basic acetate 
method, and thus freed from copper. Details of i)rocedure arc given under the 
determination of impurities in copi^er in the chapter on the subject. 

SEPARATIONS 
Isolation of Arsenic by Distillation as Arsenous Chloride ^ 

By this method arsenic may be separated from antunony, tin, and from 
other hea\y metals. It is of sjxicial value in the direct determination of arsenic 
in iron ores, copper ores, and like products and has a wide ai)plication. The 
procedure depends upon the volatility of arsenous chloride at temperatures lower 
than the other heavy metals. In a current of HCl gjus, arsenous chloride begins 
to volatilize below 108** C, and is actively volatile at 120° C; antimony 
starts to volatilize at 125** C, but is not actively volatile until a temperature 
of 180** has lx)en reached. The boiling-point of arsenous chloride, AsCU, 
is 130.2**; anthnony trichloride, SbCb, is 223.5**; and that of stannous chloride, 
SnCU, is over 603**; other chlorides having still higher boiling-points. Tin in 
its higher form, SnCU, is readily volatile, boiling-i)oint is 114** C, so that it is 
necessary to have it in its divalent form to effect a separation from arsenic. 
When heavy metals are present in the residue remaining from the arsenic dis- 
tillate, or when zinc chloride is added to raise the boiling-point, antimony 

* The ore may be brought into solution by fusion with a mixture of sodium car- 
bonate, potassium nitrate and zinc oxide, 1:1 : 2. The fusion heinp made in a platinum 
dish. The potassium iodide procedure may be followed for reduction of arsenic. (See 
Lead Arsenate.) 

* J. E. Stead's Method. R. C. Roark and C. C. McDonnell, Jour. Ind. Eng. Chcm., 
VIII, 4, 327 (1916). 



34 



ARSENIC 



may also be separated by distilktion by carrying the solution to near dryness, 
adding concentrated HCl by means of a separatory funnel, drop by drop, during 
further distillation of the concentrate. Areenic may be determined in the 
distillate (first portions) either gravimetric ally or volumetricallj'. 

Procedure. If arsenic is present as arsenic chloride, as prepared in the 
method for solution of iron ores, the sample may be transferred directly to 
the distillation flask by means of concentrated, arsenic-free hydrochloric acid. 
If a preliminary separation of other metals has been made and arsenic is 
present (along with antimony and tin) as a sulphide, it is oxidized by addition 
of concentrated HCl and sufficient potassium chlorate to cause solution and 
oxidation of fiee sulphur, and the chlorate decomposed by evaporation to 




Fia. 1. — Apparatus for the Distillation of Arsenoua Acid. 

dryness; or if preferred, by evaporation of the alkaline solution to dryneM, 
oxidation with fuming nitric and re-evaporation to dryness to expel the nitric 
acid. The residue is taken up with hydrochloric acid and washed into the 
flask with strong hydrochloric acid as directed above. 

Distillation. The sample, in a half-liter distilling flask (Fig. 1, "5") is 
made up to about 150 cc. with concentrated hydrochloric acid and about 
5 grams of cuprous chloride, Cu,Cl„ are added. The apparatus is connected up 
as shown in the illustration. Fig. 1. The end of the condenser dips into 400 
cc. of cold water in a large beaker (1 liter) or flask ("4"). The solution is 
cooled by piacuig it in ice-water or cold running water. The sample is satu- 
rated with dry hydrogen chloride gas generated by dropping concentrated 
sulphuric acid into strong hydrochloric acid ("3") and passing the gas through 



ARSENIC 35 

sulphuric acid ("1") (sp.gr. 1.84) as shown in cut. When the point of satu- 
ration is reached the gas begins to bubble through the solution instead of 
being absorbed by it. When this occurs, heat is applied and the solution brought 
to boiling, the current of HCl gas being continued. At a temperatiu^ of 108 
to 110** C. the first 100 cc. will contain practically all of the arsenic. About 
two-thirds of the solution is distilled off. It is advisable to add more hydro- 
chloric acid to the residue in the flask, together with cuprous chloride, and repeat 
the distillation into a fresh lot of water. This may be done during the esti' 
mation of arsenic in the first distillate. 

Arsenic may be determined in the distillates either gravimetrically or voh*- 
metrically. The volumetric procedures for arsenic, in this isolated form, ara 
generally to be preferred, since they are both rapid and accurate. For amounts 
over 0.5% arsenic, the iodine method is recommended, for smaller amounta 
(arsenic in crude copper), precipitation with silver nitrate and titration of tba 
silver salt is best. Exceedingly small amounts are best determined by tba 
Gutzeit method, page 40. 

Commercial hydrochloric acid invariably contains arsenic, so this must ba 
purified by redistillation in presence of an oxidizing agent to oxidize the arsenic 
to the non-volatile arsenic pentachloride, AsCU, form, (Fig. 5) or by treatment 
with KS and filtration. A blank run should be made on the reagents used* 
especially when traces of arsenic are to be determined. 

Separation of Arsenic from Antimony and Tin by Precipitation 
as Sulpliide in a Strong Hydrocliloric Acid Solution 

This procedure for isolation of arsenic depends upon the insolubility of the 
sulphide of arsenic in strong hydrochloric acid, whereas that of antimony dissolves. 
The sulphide of tin is also soluble. 

Procedure. The metals present in their lower conditions of oxidation are 
precipitated as sulphides in i)resence of dilute hydrochloric acid (5% solution) 
to free them from subsequent groups (Fe, Al, Ca, etc.). The soluble members 
of the hydrogen sulphide group are now dissolved and separated from copper, 
lead, etc., by caustic as follows: The greater part of the washed precipitate is 
transferred to a small casserole, that remaining on the filter pai)er is dissolved 
off by adding to it a little hot dilute potash solution, catching the filtrate in the 
casserole. About 5 grams weight of solid potassium hydroxide or sodium hydroxide 
is added to the precipitate. Arsenic, antimony, and tin sulphides dissolve. The 
solution is filtered if a residue remains, and the filter washed. This preliminary 
treatment is omitted if alkaline earths and alkalies are the only contaminating 
elements present. 

The casserole containing the sample is covered and placed on a steani bath. 
Chlorine is now conducted into the warm solution for an hour, whereby the alkali 
is decomposed and antimony and arsenic oxidized to their higher state. Sufficient 
hydrochloric acid is added to decomiK)se the chlorate formed, and the uncovered 
solution evaporated to half its volume. An equal volume of hydrochloric acid is 
added and the evaporation repeated, to expel the last trace of chlorine. The acid 
solution is washed into an Erlenmeyer flask, cooled by ice to 0° C. and two volumes 
of cooled, concentrated, hydrochloric acid added. H2S gas is rapidly passes into 



36 ARSENIC 

this solution for an hour and a half. The flask is now stoppered and placed in 
boiling water for an hour. The yellow arsenic sulphide, AS2S6, is filtered through 
a weighed Gooch crucible, wa.shcd with hydrochloric acid, 2:1, until free from 
antimony, i.e., the washing upon dilution remains clear. The residue is now 
washed with water, followed by alcohol, and may be dried and weighed as As^S^, 
or determined volumetrically. Antimony and tin are determined in the filtrate. 
McCay recommends washing AS2S5 with alcohol, CS^ and finally alcohol.* 

The sulphide may be dissolved in concentrated sulphuric acid by heating 
to sulphuric acid fumes and until the solution becomes clear. No arsenic is lost, 
provided the heating is not unduly prolonged. Fifteen to twenty-five minutes 
is generally suflScient to dissolve the sulphide and expel SO2, etc. The acid may 
be neutralized with ammonia or caustic, made again barely acid and then alkaline 
with bicarbonate of soda, and arsenous acid titrated with iodine.* 



GRAVIMETRIC METHODS FOR DETERMINATION OF 

ARSENIC 

As in the case of antimony, the accuracy and rapidity of the volumetric 
methods for the determination of arsenic make these generally preferable to 
the more tedious gravimetric methods. The following methods, however, are 
of value in certain analytical procedures. 

Determination of Arsenic as the Trisulphide, AS2S3 

Arsenic acid and arsenates should be reduced to the arsenous form before 
precipitation as the sulphide. The procedure is especially adapted to the 
isolation of arsenic from other elements, when this substance is present in the 
solution in appreciable quantities, advantage being taken of the extreme dif- 
ficulty with which arsenous sulphide, AsjSa, dissolves in hydrochloric acid 
solution. 

Procedure. The solution containing arsenic in the arsenioui form is made 
strongly acid with hydrochloric acid and hydrogen sulphide passed into the 
cold solution to complete saturation. The hydrogen sulphide pressure generator 
is recommended for this treatment. Figs. 3 and 4. The precipitate is filtered 
into a weighed Gooch crucible (previously dried at 105° C), the compound 
dried at 105° C. to constant weight and weighed as AS2S3. 

Factors. AS2S8 X 0.6091 = grams As. 

AsaS, X 0.8042 = grams AsjO,. 
AsjOs X 1 . 1 61 6 = grams As20j. 
As20fcXl.3134 =grams HaAsO^JHjO. 
AssSsX 1 .2606 = grams AszSj. 

Note. Arsenic may also be determined as arsenic sulphide by passing a rapid 
stream of HaS into a cooled solution of arsenics acid containing at least two parts of 
concentrated hydrochloric acid for each j)art of water present in the solution. 

» T/e Roy W. McCay, Chem. News, 66, 262 (1887). 

* J. and H. S. Pattinson, Jour. 80c. Chem. Ind., 1898, p. 211. 



Determinationof Arsenic as Magnesium Pyroarsenate 

The method worked out by Levol depends upon the precipitation of arsenic 
as MgXH,AsOi-6H,0, when maRnesia mixture is added to an ammoniacal 
solution of the arsenate. Although 600 parts of wat«r dissolve 1 part of the salt, 
it in practically ins<)luljle in a 2i per cent ammonia solution, 1 part of the anhydrous 
salt rcquirinK 24,558 parts of the ammonia water accordini; to Virgili.' The 
cempound loses 5J molecules of water at 102° C. and all of the water when 




Fin. 2. — l"rbasch'» Hydrogen Sulphide Generator. 



.0)M.1MO. Tb<' Analyst (1010) U, 5S8). 
. naluiuti^ aquwu* Mlutlon to 1» obtained. 
« <ulpbi.l<' IK pHvcd in III. The hyilroiKn 
■ ■ *^ ■ ■ ■ WalPT is placBd 
n A. Hyarogen 



in f and 11. U gBH i> trauiml Iho tap* A an.l D an: oprnc-d and Hi.-5 di 

■iilphidr *at«[ in obtainij by opeainR th? pincli cock <7 of Ihc buTvtte. tl 

•imuKidHualy rrplaci^ from Ibi- vpwrl II. Tbg conlainer ia madi: of dar.-..^....^ „u.^. w |..i>.<^> 

thr hydroKcn milpiiidB wBier [lom ligbt. WadT may hv drawn Into II, wben required by opening 

tlv pinch CDcli leading to tbe bottio I. 

strongly ignited, forming in presence of oxj'gen the stable magnesium pyro- 
arsenate, MgiAsiO?, in which form arsenic is determined. 

Procedure. The solution containing the arwnic, in the form of arsenate, 
and having a volume not exceeding 100 ec. jwr 0.1 gram arsenic present, is treated 
with 5 cc. of concentrated hydrochloric acid, added, with constant stirring, 
drop by drop. Ten cc. of magnof^ia mixture arc added (Reagent ^.W grama 
MgCI,+70 grams NH*CI-|-fi;W cc. H,0 and made up to 1000 cc. willi XH.OH, 
Hp.gr. O.iKi), for each 0.1 gram of arsenic present. Ammonia solution (sp.gr. 0.96) 
L-< added from a burette, with stirring, unlil the mixture is neutrali;!cd (a red 
color imparted to the solution in prcseix-e of plicnolphtlmleiii indicator), and 
' Average of three results. J. F. Virgili, Z. annl. Chem., 44, 504 (1905). 



38 



AESENIC 



then ammonia added in excess equal to one-third the volume of the neutralized 
solution. The precipitate is allowed to settle at least twelve hours and is then 
filtered into a weighed Gooch crucible and washed with 2.5% ammonia until 
free from chloride. After draining as completely as possible by suction the 




cock . 



Mercuru Vafve 

and PfessuKf 

Gauge 



RybberBulb 
wifhAiflbhts 



Fig. 3. — Scott's Hydrogen Sulphide Generator. 

Fig. 3 ibows a oonvenient form of a generator for obtainins hydrogen sulphide gas under pres- 
sure. The apparatUB is the writer's modification of the Banks* generator shown in Fig. 4. and is 
designed for large quantities of hydrogen sulphide gas. The cylinder A A"\^ constricted, as shown, to 
support perforated lead disk G, upon which rests the iron sulphide. The lower end of the cham- 
ber is closed to catch small particles of FeS that may be carried through the perforations of the 
disk. Small openings admit the acid to A'. The level of the acid is below the disk G, so that the 
acid only comee in contact with the sulphide when pressure is applied by means of the rubber bulb 
E^ the stopcock S* being open and 5' closed. The mercuiy gauge C is adjusted to blow out at a 
given pressure, to prevent accident, the bulb D preventing the mercury from being blown out of the 
apparatus. A small opening in D allows the escape of the gas. When the apparatus is in opera • 
tion» H is connected to an empty heavv-walled bottle, which in turn is attached with glass tube 
connection to the pressure flask in which the precipitation of the sulphide is made, the flask being 
dosed to the outside air. B}r pressure on the rubber bulb E, acid is forced into the chamber A 
past the disk into the sulphide in A. The entire system will now be under the pressure indicated by 
the gauge C The pressure is released by opening the stopcock 8' and the flask containing the pre- 
cipitate then disconnected. The reservoir is designed to hold about two liters of acid, and the 
cylinder containing the sulphide is of sufficient capacity to hold over one pound of FeS, so that the 
apparatus will deliver a large quantity of hydrogen sulphide. 



precipitate is dried at 100** and then heated to a dull red heat (400 to 500** C), 
preferably in an electric oven, until free of ammonia. The temperature is then 
raised to a bright red heat (800 to 900° C.) for about ten minutes, the crucible 
then cooled in a desiccator and the residue weighed as Mg^AsaOv. 

Factors, MgjAsiO; X 0.4827 = As, or X 0.6373 ^AsjOi, or X0.7403=As2O., 
or X 0.7925 -AsiSi. 



Notes. Iq place of an electric furnace the Gooch crucible may oe placed in 
a lai^er nco-perforated crucible, (he bottom of the Gooch being 2-3 mm. above the 
bottom of the outer crucible. The product may now be heated in presence of a 
current of oxygen pawed through a perforation in the covering lid of (be Qooch, or 




Fia. 4. — Banks' Hydrogen Sulphide Generator. 

FIc- 4 ahom k ample and cflcdive prruure ggaentor, dpagnei) by Bulu. The ttMntion at 

.. .. .».-, ;._, ;. ^_:,.. ._ .U ,.-,_ n H i, „peri,|fy Bd«pl«l 

^ mm. ir»ic isoiviaiuii •DDuaiiu is ae«rea lor nuiWDU id — ■--■■-- .- j . 



irhBic isdividukl appui 



r nudeati id qiuliutiva uid quuiUtft- 



in place of the oxyoen. a thin layer of powdered NHtNOj may be placed on the 
aR>cnate residue ana the heat gradually applied until the outer crucible attains a 
light red glow. 

VOLUMETRIC METHODS FOR THE DETERMINATION OF 

ARSENIC 

Oxidation of the Arsenous Acid with Standard Iodine > 

This procedure is applicable for the determination of arsenic in acids, after 
reduction of arsenic to its arsenous form, for valuation of arsenic in the tri- 
ozide, for detemuDation of arsenic isolated by distillation as arsenous chloride, 
for arsenic in arseoites and reduced arsenates in insecticides, etc. The method 
depends upon the reaction— As,0,+2H,0+2It=As,0,+4HI. The hberated 
hydriodic acid is neutralized by sodium bicarbonate. The trace of excess 
iodine is detected by means of starch, a blue color being produced. 

Procedure. If tbesoJutioi) in ncid, it U neutralized by sodium or potassium 
hydmxitie or carbonate (phoriolplitlmlpin indicator) then made slightly acid. If 
the solution is alkaline, it is made slightly ncid. Two to 3 grams of sodium 
bicarlM>nate are added together with starch indicator and the solution titrated 
with tenth normal i<Kline solution, the iodine Iteing added cautiuuijly from a 
burette until a permanent blue color develops. 

> Mohr's Method. 



40 ARSENIC 

One cc. N/10 iodine =0.00375 gram As, or 0.004948 gram AsjOa. 
ASiOzX 1.1616 = AsA. AsX 1.3201 =As203 or X 1.5336 =AsA. 
As2OaX0.7575=As. 

Volumetric Determination of Arsenic by Precipitation as Silver 

Arsenate 

Bennett's modification of Pearce's method, combining Volhard's, depends 
upon precipitation of arsenic, from a solution neutralized with acetic, by addition 
of neutral silver nitrate solution; the silver arsenate is dissolved in nitric acid, 
and the silver titrated with standard thiocyanate. 

Procedure. 0.5 gram, or less, of the finely powdered substance is fused 
with 3 to 5 grams of a mixture of sodium carbonate and potassium nitrate (1:1) 
about one-third being used on top of the charge. The cooled mass is extracted 
with boiling water and filtered. The filtrate, containing the alkali arsenate, 
is strongly acidified with acetic acid, boiled to expel the carbon dioxide, 
then cooled and treated with sufficient sodiuni hydroxide solution to give an 
alkaline reaction to phenolphthalein indi(;ator. The purple red color is now 
discharged from the solution ])y addition of acetic acid. A slight excess of 
neutral silver nitrate is vigorously stirred in and the precipitate allowed to settle 
in the dark. The supernatant liquid is poured off tlu-ough a filter and the 
precipitate washed by decantation with cold distilled water, then thrown on the 
filter and washed free of silver nitrate solution. The funnel is filled with water 
and 20 cc. of strong nitric acid added. The dissolved silver arsenate is caught 
in the original beaker irt which the precipitation was made, the residue on the 
filter washed thorouglily with cold water and the filtrate and washings made 
up to 100 cc. The silver is now titrated by addition of standard anunonium 
or potassium thiocyanate, until a faint red color is evident, using ferric ammonium 
alum indicator, according io the procedure described for determination of silver. 
(See Chlorme and Silver Chapters.) 

One cc. N/10 thiocyanate =0.010788 grain Ag. 
Factor. AgXO.2316 =As. 

Note. The silver arsenate salt is nearly six times the weight of arsenic, so 
that very small amounts of arsenic may l)e detennined by the procedure, hence 
it is not necessary to use over 0.5 gram of the material. For traces of arsenic the 
Gutzeit method, following, should be used. 

DETERMINATION OF SMALL AMOUNTS OF ARSENIC 

Modified Qutzeit Method ^ 

The following procedure furnishes a rapid and accurate method for deter- 
mination of exceedingly small amounts of arsenic ranging fnim 0.001 milli- 
gram to 0.5 milligram AS2O6. It is more sensitive and less tedious than the 
Marsh test. The details, given below with slight modifications, have been 
carefully worked out in the lai)orat()ries c>f the (i(»neral Chemical Company * 
and have proved exceedingly valuable hi esthnating small amounts of arsenic 
in acids, bases, salts, soluble arsenic in load arsenate and zinc arsenite and other 
insecticides, traces of arsenic in food product**, baking powders, caimed goods, etc. 

^ Paper by W. S. Allen and R. M. Palmer General Chemical Company, pre- 
sented before the Eighth International Congress of Applied, Chemistry'. 



ARSENIC 41 

The method depends upon the evolution of arsine by the action of hydrogen 
on arsenic compounds under the catalj'tic action of zinc, the reaction taking 
place either in alkaline or acid solutions. The evolved arsine reacts with mer- 
curic chloride, forming a colored compound. From the length and intensity 
of the color stain the amount of arsenic is estimated by comparison with 
standard stains. 

Arsine is evolved from an acid solution under definite conditions of aciditv, 
amount of zinc used, temperature, stnmgth of solution of mercuric chloride 
used in sensitizing the test-paper, size of apparatus, volume of solution, amount 
of iron accelerator, and of stannous chloride reducer, etc., conditions wliich 
have ])roven, by extended tests, to be most efficient. These conditions must 
be adhered to for reliable results. For example, variatio n of acidity and the 
amount of zinc will produce stams of variable length and intensity with eqUal 
ajnounts of arse nicTTli^jltuh' jiemg lo nger and jess intense with tlie more rapid 
evolution of the gas. Likewise the greater the coiicentratioii of mercuric 
chloride on the sensitized paper, the sEorter the length of the stain and the 
deeper its color. 

Special Reagents. Standard Arsenic Solution, One gram of resublimed 
arsenous acid, As^Oa, is dissolved in 25 cc. of 20% sodium hydroxide solution 
(arsenic-free) and neutralized with dilute sulphuric acid. This is diluted with 
fresh dl^^tilled water, to which 10 cc. of 95% H1SO4 has been added, to a 
volume of 1000 cc. Ten cc. of this solution is again diluted to a liter with dis- 
tilled water containing acid. Finally 100 cc. of the latter solution is diluted 
to a liter with distille<l water containing acid. One cc. of the final solution 
contains 0.001 milligram AsiOi. 

Standard Stains. Two sets of stains are made, one for the small apparatus 
for detennining amounts of AsjOj ranging fn)in 0.001 to 0.02 milligram, and a 
second set for the larger-sized apparatus for determining 0.02 to 0.5 milligram 
As^Oa. Stains made by AsiOi in the following amounts are convenient for the 
standard sets; e.g., small apparatus, 0.001, 0.002, 0.004, 0.006, 0.01, 0.15, 0.02 
milligram As,Oa. Large apparatus, 0.02, 0.05, 0.1, 0.2, 0.3, 0.4, 0.5 milligram 

ASjOa. 

In making the stain the requbito amount of standard reagent, AsjOs solution, 
is placed in the Gutzeit bottle with the amounts of reagents prescribed for the 
regular tests and the run made exactly as prescribed in the regular procedure. 

Preservation of the Stains. The strips of sensitized pai)er with the arsenic 
stain are dipped in molten paraffine (free from water), and mounted on a sheet 
of white paper, folded l)ack to fonn a cylinder. The tube is placed in a 
glass test-tube containing phosphorus pentoxide, which Ls then closed u^ 
a stopper. It is important to keep the stained strip dr>', otherwise the 
stain soon fades, hence the i)aper on which the strips are mounted and the 
glass test-tul>e, etc., nmst be i)erfectly dr>'. It is advisable to ket»p the standard 
in a hydrometer case, while not in use, as light will gradually fade the color. 

Sensitized Mercuric Chloride (or Bromide) Paper. 20X 20 in. Swedish Filter 
Paper No. is cut into four equal squares. For use in the large Gutzeit appa- 
ratus the paper is dipped into a 3.5% solution of mercuric chloride (mercuric 
bromide may be used in place of the chloride) or if it is to be used in the small 
Gutzeit apparatus it is dipped into a 0.5% mercuric chloride solution. (The 
^-caker the solution, the longer and less intense wull be the stain.) A squeegee 
roller is passed gently over the impregnated paper, placed on plate glass, to 



42 ARSENIC 

remove the greater part of the reagent. The paper, placed on a clean dry 
cloth, is dried at a temperature of 100° C. Immediately upon drying it is 
removed from the oven (HgC'li is Blightly volatile witJi heat), half an inch of 
the outer edge trimmed off, since this Is apt to contain more of the reagent, 
and the paper cut into strips. The paper with more concentrated reagent 
is cut into strips 13 cm. by 5 mm. and that with 0.5% mercuric chloride 
into strips 7 cm. by 4 mm. The paper is preserved in bottles with tight^fitting 












Fio. 6.— Purification o( Hydrochloric Acid. 

stoppers. Standards should be ma^p with each batch of paper. Paper with 
a white deposit of HgC'li should not be used. 

Ferric Ammonium Alum. ElRhty-four grams of the akim with 10 cc. of 
mixed acid is dissolved and made up to a liter. Ten cc. of this solution con- 
tains approximately 0,.t gram FciOi. 

Lead Acetate. One per cent solution with sufficient acetic acid to clear 
the .■K>lution. 

Zinc. Arsenic-free zinc shot, 3 to 6-in. mceh. The sine is treated with C. P. 
hydrochloric acid, until the surface of the unc becomes clean and dull. It is 



ARSENIC 43 

then washed, and kept, in a casserole, covered with distilled water, a clock-glass 
keeping out the dust. 

Mixed Acid One volume of arsenic-free H2SO4 is diluted with four vol- 
umes of pure water and to this are added 10 grams of NaCl per each 100 cc. of 
solution. 

StannouB Chloride, Eighty grams of stannous chloride disisolved in 100 cc 
of water containing 5 cc. arsenic-free hydrochloric acid (1.2 sp.gr.). 

ArteniC'free Hydrochloric Acid. The commercial acid is treated with ]X)tas- 
sium chlorate to oxidize the arsenic to its higher form and the acid distilled. 
The distilling apparatus may be arranged so that a constant distillation takes 
place, acid from a large container dropping slowly into a retort containing 
potassium^ chlorate, fresh hydrochloric acid being supplied as rapidly as the 
acid distills. See Fig. 5 on page 42. 

Lead Acetate Test Paper for Removal of HtS. Large sheets of qualitative 
filter paper are soaked in a dilute solution of lead acetate and dried. The paper 
is cut into strips 7X5 cm. 

Blanks should be run on all reagents used for this work. The reagents are 
arsenic-free if no stain is produced on mercuric chloride paper after forty-five 
minutes' test. 

Special Apparatus. The illustration, Fig. 6 (page 46), shows the 
Gutzeit api)aratus connected up, ready for the test. The dimensions on the 
left-hand side are for the small apparatus and those on the right for the large 
form. Rubber stoppers connect the tubes to the bottle. The apparatus con- 
sists of a wide-mouth 2-oz. or 8-oz. bottle according to whether the small or 
large apparatus is desired, two short glass tubes containing dry lead acetate 
paper and moist glass wool for removal of traces of hydrogen sulphide and a 
sDoiJEtll-bore tube containing the strip of mercuric chloride paper. 

Preparation of the Sample 

The initial treatment of the sample is of vital importance to the Gutzeit 
Method for determining traces of arsenic. The following procedures cover 
the more im]X)rtant materials or substances in which the chemist will be called 
upon to determine minute amounts of arsenic. 

Traces of Arsenic in Acids. The acid placed in the Gutzeit apparatus 
should be equivalent to 4.2 grams of sulphuric acid or 3.1 grams of hydrochloric 
acid and should contain 0.05 to 0.1 gram FezOi equivalent. If large samples 
are required for obtaining the test it is necessary either to expel a portion of 
the acid in order to obtain the above acidity or tx) make standard stains under 
similar conditions of acidity. It must be remembered that arsenous chloride 
is readily volatile, whereas the arsenic chloride is not, hence it is necessary to 
03ddize arsenic before attempting to expel acids. If nitric acid or bromine or 
chlorine (chlorate) be added for this purpose, !t must be expelled before attempting 
the Gutzeit test. Nitric acid may be expelled by adding sulphuric acid and 
taking to SOi fumes. Free chlorme, bromine, or iodine will volatilize on 
warming the solution. Chlorine in a chlorate is expelled by taking the sample 
to near drjmess in presence of free acid. Sulphurous acid or hydrogen sul- 
phide, if present, should be oxidized by addition of potassium permanganate 
in slight excess (faint pink color) the excess destroyed by a drop or so of oxalic 
acid. SOs is reduced by zinc and hydrogen to HS, which forms black HgS 



44 ARSENIC 

with mercuric chloride, hence oxidation of SOj and H2S are necessary before 
running the test. 

Sulphuric Acid. With amounts of arsenic exceeding 0.00005% AssOj, 5 
to 10 grams of acid, according to its strength, are taken for analysis and diluted 
to 15 or 20 cc. If H-S or SO2 arc present, expo! by boiling for fifteen or twenty 
minutes. No loss of arsenic occurs, oven though the acid begins to fume; however, 
prolonged fuming should be avoided. In mixed acid containing nitric acid the 
sample is taken to SO3 fumes to expel nitric acid. The procedure given later for 
the regular determination is now followed. 

For estimating very minute amounts of arsenic, 0.000005 to 0.00005% As20», 
it is necessary to take a 25- to 50-gram sample for analysis. The acid is treated 
as directed above for removal of HjS or SOj or nitric acid and diluted in the 
Gutzeit apparatus to at least 130 cc, using the large apparatus. Upon the 
addition of iron and stannous chloride as directed in the procedure the run 
is made with 25 pieces of zinc. The stains are compared with standard stains 
produced by known amounts of arsenic added to 50-gram portions of arsenic- 
free sulphuric acid of strength equal to that of the sample. The stains are longer 
and less intense than those produced by less acid. 

Hydrochloric Acid. Twenty cc. Is taken for analysis (sp.gr. being known); 
the sample should contain an acid equivalent of about 3.1 grams of hydrochloric 
acid. Chlorine is expelled by bubbling air through the acid before taking a 
sample. The procedure is given for further treatment of the sample following 
the section on preparation of the sample. v 

Nitric Add. One hundred cc. of the acid (sp.gr. being known) is evap- 
orated with 5 cc. of concentrated sulphuric acid to SOs fumes, to expel nitric 
acid. Arsenic is determined in the residue by the standard procedure. 

Iron Ores, Pyrites, Burnt Pyrites, Cinders, etc. One gram of the finely 
ground ore is oxidized by treating with 5 cc. of a mixture of 2 parts liquid 
bromine and 3 parts of carbon tetrachloride. After fifteen minutes, 10 cc. of 
concentrated nitric acid are added and the mixture taken to dr>^ness.. Five cc. 
of concentrated sulphuric acid (95%) are added and the mixture taken to SOi 
fumes to expel the nitric acid. The cooled sample Ls taken up with 50 cc. of 
water and digested until all of the iron sulphate has dissolved; it is now washed 
into a 100-cc. flask, made to volume, and arsenic determined in an«aliquot portion 
in the usual way, given later. Insoluble FejOs, briquettes, etc., is best dis- 
solved by fusion with potassium bisuiphate, KHSO4. The fused mass is dis- 
solved in warm dilute hydrochloric acid, and then washed into the Gutzeit 
bottle. 

Alumina Ores. Bauxite. One gram of bauxite is treated with one part 
of concentrated nitric acid and 6 parts of concentrated hydrochloric acid, and 
taken to dryness on the water bath. The residue is taken up with an equiv- 
alent of 4.7 grams of hydrochloric acid or 6.3 grams of sulphuric acid in a 
volume of 25 cc. and the mix heated until the material has dissolved. The 
sample is diluted to exactly 100 cc. and arsenic detemimed on an aliquot 
portion. 

Phosphates, Phosphoric Acid. Arsenic, in phosphoric acid, combined or 
free, cannot be determined in the usual way, as PjOs has a retarding effect 
upon the evolution of arsine, so that the results are invariably low, and snmll 
amounts of arsenic escaping detection. Arsenic, however, may be volatilized from 
phosphates and phosphoric acid, as arsenous chloride, AsCU, in a current of 



ARSENIC 45 

hydrogen chloride by heating to boiling. One gram or more of the phosphate 
is placed in a small distilling flask, connected directly to a 6-in. coil condenser 
dipping into the Gutzeit bottle, containing 20 to 30 cc. of cold distilled water. 
A second bottle connected in series may be attached for safeguarding loss (this 
seldom occurs). Fifty cc. of concentrated hydrocliloric acid are added to the 
sample and 5 grams of cuprous chloride. Arsenic is distilled into the Gutzeit 
bottle by heating the solution to boiling and passing a current of air through 
strong hydrochloric acid into the distilling flask by applying suction at the re- 
ceiving end of the system. All of the arsenic will be found in the first 10 or 15 
cc. of the distillate. Arsenic may now be evolved after addition of iron, stannous 
chloride and zinc, as directed in the procedure. 

Salts, Sodium Chloride, Magnesium Sulphate, etc. One-gram samples 
are taken and dissolved in a little water and an equivalent of 6.3 grams 
of sulphuric acid added. The solution of iron and stannous chloride having 
been added, the run is made with 25 pieces of zinc, placed in the Gutzeit 
bottle. 

Baking Powder, Other than Phosphate Baking Powder. A ten-gram 
sample is heated with 10 cc. hydrochloric acid, 1-0 cc. of ferric ammonium alum 
and 30 cc. of distilled water, until the starch hydrolyzes. 0.5 cc. of stannous 
chloride is added to the hot solution and the mixture washed into the Gutzeit 
apparatus. Twenty-five pieces of zinc are added and the arsenic determined as 
usual. 

Phosphate Baking Powders. Ten grams of the material mixed to a paste 
with about 50 cc. of hydrochloric acid are transferred to a small distilling flask 
with a few cc. of HCl. A tube, connected to a bottle of strong hydrochloric 
acid, passes into the mixture in the flask through a ground glass stopper. The 
flask is attached to a tube, which dips into water in a Gutzeit bottle. Two 
grams of cuprous chloride are added, the apparatus made tight and the flask 
immersed in boiling hot water. By aspirating air through the system into the 
Gutzeit bottle, which is water cooled, arsenic distills into the bottle and may be 
determined by the procedure outlined. 

Arsenic in Organic Matter, Canned Goods, Meat, etc. The flnely chopped, 
well-mixed sample is placed in a large flask and enough water added to pro- 
duce a fluid mass. An equal quantity of concentrated hydrochloric acid and 
1 to 2 grams of potassium chlorate are added. The flask is shaken to mix the 
material and it is then placed on the steam bath. Upon becoming hot, nascent 
chlorine is evolved and vigorously attacks the organic matter. Half-gram 
portions of ]X)tassium chlorate are added at five-minute intervals, shaking the 
flask frequently. When the organic material has decomposed and the solution 
becomes a pale yellow color, the mass is diluted with water and filtered. Arsenic 
will be f ound in the filtrat e. A white; amorphous substance generally remains 
on the filter, when cadaver is being examined. The filtrate is diluted to a 
given volume and an aliquot portion taken for analysis. This is evaporated to 
near dryness to expel excess of acid and decompose chlorates. "/An equivalent 
of 4.7 grams of hydrochloric acid is added (three times this amount for the 
laiige apparatus), the volume of the solution made to about 30 cc, 10 cc. of ferric 
ammonium alum and 0.5 cc. of stannous chloride added, and the solution poured 
into the Gutzeit apparatus for the test as given below. 



46 



ARSENIC 



-* 



1.5 mm. Dong 
Constricted 
12 cm. from 
Upper End. 



Procedure for Making the Test 

For amounts of arsenic var^^ing from 0.001 milligram to 0.02 milligram 
AssOs, the small apparatus is used. The vol\>me of the solution should be 50 cc. 
It should contain an equivalent of 4.2 to 6.3 grams sulphuric acid and should 

have about 0.1 gram equivalent of F^t re- 

Loroe Apparatus ^^^^^ ^X ^.5 cc. of Stannous chloride solution. 
Arsine is generated by adding 12 to 25 pieceSs of 
zinc. 

For amounts ranging from 0.02 to 0.5 milli- 
gram AssOi, the large apparatus is used. The 
volume of the solution should be about 200 cc. 
and should contain an equivalent of 18.5 grams 
of sulphuric acid and should have O.r gram 
equivalent of FcjOs, reduced by 0.5 cc. stannous 
chloride solution. Arsine is generated by adding 
25 to 40 pieces of arsenic-free zinc. The tem- 
perature should be 105° F. The sample taken 
should be of such size that a stain is obtained 
equivalent to that given by 0.1 to 0.5 milligram 
AsiOj. 

Lead acetate paper is placed in the tube next 
to the bottle; the middle tube (see Fig. 6) con- 
tains glass wool moistened with lead acetate 
solution; the upper tube contains the test strip 
of jTiercuric chloride paper. Immediately uix)h 
adding the required amount of zinc to the solu- 
tion in the bottles, the connected tubes are put 
in position, as shown in the illustration, and 
the bottle gently shaken and allowed to stand 
for thirty minutes for the small apparatus, forty 
minutes for the large. The test paper is removed, 
dipped in molten paraffine and compared with the 
standard stains. See Plate I. 

Estimation of Per cent. 



Small Apparatus 

!0 cm. long A 
▼ mm. Bore 
Conztn'cted 
6cm. from 
Upper End. 



Strip H^i 
^perin Tube 



Scm.X 1.25cm. 

Moistened ''^\ 
with Lead 
Aceiofe 
Solution 

7cm. X 1, 25 cm. \ 
dry Lead 



6 cm. X 1,5 cm. 



9 cm X /. 5 cm. 
Acetate Fbper. 



2'02. voltfe 
60cc., ^r 
Tests of 

below 



• »■»■» 



8-oz.Boitk 
250 cc, for 
Tests of 
AS,0» 



aofmg. \yj,h::\0y'ra05mg. 

A5,o; ■•*-•••'■ 



-=>1^^-^ 



FlQ. 6. — Gutzeit Apparatus for 
Arsenic Detenmnation. 



The milligram As^Oa staiaX 100 _ (w » n 
Weight of sample taken ' 



Notes. Ferrous iron prevents polarization.*betwccn zinc and the acid and hence 
aids in the evolution of arsmc. 

In the analysis of baking powders, bauxite, sodium or similar salts, additional 
sine — ^25 pieces, and additional acid, 0.3 grams sulphuric ucid equivalent, arc required. 

Hydrochloric acid is used in place of sulphuiic acid in cases where complete 
solution by the latter acid cannot be efTcrted. 

Standards and samples should be run under similar conditions, temperature^ 
acidity, amount of zinc, volume of solution, etc. 



v.. 



ARSENIC 47 

METHOD FOR ANALYSIS OF COMMERCIAL "ARSENIC,** 

ARSENOUS OXIDE, AS2O3 

The following constituents may be commonly present as impurities, SiOi, 
SbjO,, Fe,0„ NiO, Ck)0, CaO, SO,, Cu, Pb, and Zn. 

Determination of Moisture 

Two 10-gram samples are dried to constant weight in the oven at 100^ C- 
Loss in weight » moisture. 

Sulphuric Acid, H2SO4 

The samples from the moisture determination are dissolved in concentrated 
hydrochloric acid, heating to boiling if necessary, and the samples diluted to 
300 to 400 cc. Barium chloride solution is added in slight excess to the hot 
solution, the precipitate, BaSOi, allowed to settle and filtered and the sulphate 
dried and ignited as usual. 

BaSOiX 0.343 =S0,. 

Determination of Arsenic as AS2O3 

Duplicate 5-gram samples are dissolved in 20 grams ]X)tassium carbonate 
in 60 cc. of hot water, by boiling until solution is effected. The samples are 
made up to 1 liter and aliquots of 100 cc. ( =0.5 gram) taken for analysis. The 
solution is made faintly acid with hydrochloric acid, testing the solution with 
litmus paper or by adding methyl orange directly to the solution. An excess 
of bicarbonate is added and the arsenic titrated with tenth-normal iodine 
according to the standard procedure for arsenic. One cc. N/10 I =.004948 
gram AstOi. 

Residue upon Sublimation of AS2O3. Si02, Pb, Cu, Fe203, NiO, 

C0O9 Zn 

Two 5-gram samples are weighed into tared porcelain crucibles and heated 
gently on sand baths with the sand banked carefully around the crucible so 
as to heat the entire receptacle. After the greater part of the arsenous oxide 
has volatilized, the crucible is ignited directly in the flame to a dull red heat, 
until fumes are no longer given off. The residue is weighed as total non-sub- 
limable residue. 

Silica 

The residues are transferred to beakers and treated with aqua regia, taken 
to drjmess, and the silica dehydrated at 110® C. for an hour or more. The 
residue is taken up with hot dilute hydrochloric acid, boiled, and the silica 
filtered off, ignited, and weighed. 

Lead and Copper 

The filtrate from the silica is " gassed " with HiS and the precipitate filtered 
off. The filtrate is put aside for determination of iron, etc. The precipitate is 
dissolved in hot dilute nitric acid, 2 to 3 cc. of concentrated sulphuric acid added, 



48 ARSENIC 

the solution taken to SO, fumes, the cooled concentrate diluted to 20 or 30 cc, 
and the lead sulphate filtered off, ignited, and weighed as PbS04. 

The filtrate from the lead sulphate containing the copper is treated with 
aluminum powder and the copper thrown out of solution ; the excess of alumi- 
num is dissolved with a few cc. of hydrochloric acid. The filtrate should be 
tested for copper with H^S and the precipitate added to the copper thrown 
out by the aluminum. The copper on the filter is dissolved in hot dilute nitric 
acid, the extract evaporated to 2 or 3 cc, the acid neutralized with ammonia 
and then made acid with acetic, potassium iodide added and the liberated iodine 
titrated with standard thiosulphate solution according to the regular scheme for 
copper. 

Iron, Nickel, Cobalt, and Zinc 

The filtrate from the H2S Group is boiled to expel the HjS and the iron 
oxidized by addition of nitric acid and boihng. The iron (and alumina) is 
precipitated with ammonium hydroxide and the precipitate filtered off and 
washed several times with hot water. If alumina is suspected (light-<;olored 
precipitate) it may be determined by the difference method — ignition of the 
precipitate, weighing, and finally subtracting the iron found by titration with 
standard stannous chloride solution. The iron is dissolved in hydrochloric acid 
and titrated hot with stannous chloride solution. 

The filtrate from the iron is boiled and a 1% alcoholic solution of dimethyl- 
glyoxime added to precipitate the nickel. The salt is filtered on a tared 
Gooch, the precipitate dried at 100° C, and weighed. The weight of the 
salt X 0.2032 =Ni. 

The filtrate from the nickel is boiled until all the alcohol has been driven 
off and the cobalt precipitated by addition of sodium hydroxide in excess 
filtered, ignited, and weighed as CoO. 

The filtrate is made acid with hydrochloric acid, and then alkaline with 
ammonium hydroxide and colorless sodium sulphide solution added to pre- 
cipitate the zinc. The mixture is boiled five to ten minutes, the precipitated 
ZnS allowed to settle, filtered off, and waslicd once or twice and then dissolved 
in hydrochloric acid and the zinc determined by titration directly with potassium 
ferrocyanide, or by converting to the carbonate by addition of potassium car- 
bonate, filtered and washed free of alkali, the precipitate dissolved in a known 
amount of standard acid, and the excess acid titrated with standard caustic 
(methyl orange indicator) according to the procedure given for zinc. 
H,SO4X0.06665=Zn. 

Antimony and Calcium Oxides 

Two 15-gram samples are treated with 300 cc. of concentrated hydrochloric 
acid, boiled down to 50 cc. to expel the arsenic as AsCla, an ecjual amount 
of concentrated hydrochloric acid is added, and the last traces of arsenic pre- 
cipitated by H2S passed into the hot concentrated hydrochloric acid solution. 
The arsenous sulphide, AsjSs, is filtered off. Antimony is precipitated by dilut- 
ing the solution with an ecjual volume of water, the solution having been concen- 
trated by boiling down to about 50 cc. The SbjSs is filtered off, washed several 
times with hot water, dissolved by washing through the filter with concentrated 
hydrochloric acid, and antimony detenuined in the strong hydrochloric acid 
solution by the potassium bromate method — addition of methyl orange indicator 



ARSENIC 49 

and titration with standard potassium bromate added to the hot solution to the 
disappearance of the pink color of the indicator. 

The filtrate from the antimony is concentrated, made slightly alkaline with 
ammonium hydroxide, and gased with hydrogen sulphide to remove iron, nickel, 
cobalt, zinc, chromium, and last traces of lead, etc. The filtrate is then con- 
centrated and made acid with crystals of oxalic acid, boiled and methyl orange 
added and then ammonia drop by drop slowly until the indicator changes to 
an orailge color. An excess of anunonium oxalate is now added and the beaker 
placed on the steam bath until the calcium oxalate has settled. The lime is 
now determined by filtering off the precipitate and washing, dr>'ing and igniting 
to CaO, or by titration with standard permanganate, according to the regular 
procedure for calcium. 

The author wishes to acknowledge the assistance received from Mr. J. P. Kelly and 
Dr. F. E. Hale by review and criticism of this chapter. 



BARIUM 

Wilfred W. Scx)tt 

Ba, at.wt. 137^7; sp.gr. 3.78; m.p. 850"" C; volatile at 950'' C; qxidet, 

BaO, BaOs. 

DETECTION 

Barium is precipitated as the carbonate together with strontium and cal- 
cium, by addition of ammonium hydroxide and ammonium carbonate to the 
filtrate of the ammonium sulphide group. It is separated from strontium 
and calcium by precipitation as yellow barium chromate, BaCr04, from a slightly 
acetic acid solution. 

Saturated solutions of calcium or strontium sulphates precipitate white 
barium sulphate, BaS04, from its chloride or nitrate or acetate solution, barium 
sulphate being the least soluble of the alkaline earth sulphates. 

Soluble chromates precipitate yellow barium chromate from its neutral 
or slightly acetic acid solution, insoluble in water, moderately soluble in chromic 
acid, soluble in hydrochloric or nitric acid. 

Fluosilicic acid, HsSiFe, precipitates white, crystalline barium fluosilicate, 
BaSiFe, sparingly soluble in acetic acid, insoluble in alcohol. (The fluosilicates 
of calcium and strontium are soluble.) 

Flame. Barium compounds color the flame yellowish green, which appears 
blue through green glass. 

Spectrum.^ Three characteristic green bands (a, ^, 7). 

Barium sulphate is precipitated by addition of a soluble sulphate to a solu- 
tion of a barium salt. The compound is extremely insoluble in water and 
in dilute acids (soluble in hot concentrated sulphuric acid). The sulphate is 
readily distinguished from lead sulphate by the fact that the latter is soluble 
in ammonium salts, whereas barium sulphate is practically insoluble. 

ESTIMATION 

The determination of barium is required in the. valuation of its ores, barite» 
heavy spar, BaS04; witherite, BaCOs; baryto calcite, BaCOaCaCOj. It is de- 
termined in certain white mixed paints and colored pigments, Venetian, Ham- 
burg or Dutch whites, chrome paints, etc. In analysis of Paris green, baryta 
insecticides, putty, asphalt, dressings and pavement surfacings. It may be 
found as an adulterant in foods, wood preservatives, filler in rubber, rope, 
fabrics. It is determined in salts of barium. The nitrate is used in pyro- 
techny, in mixtures for green fire. 

Preparation and Solution of the Sample 

Comjwunds of barium, with the exception of the sulphate, BaS04, are sol- 
uble in hydrochloric and nitric acids. The sulphate is soluble in hot concen- 

* See Preliminary Tests under Separations. 

50 









• • • 



9, 



• • 









• • 



• • 



• • • 

• • • 



BARIUM 51 

trated sulphuric acid, but -is reprecipitated upon dilution of the solution. The 
sulphate is best fused with sodium carbonate, which transposes the compound 
to bfoium carbonate; sodium sulphate may now be leached out with water and 
the residue, BaCOs, then dissolved in hydrochloric acid. 

• Solution of Ores. Sulphates. 0.5 to 1 gram of the finely divided ore is 
fused with 3 to 5 grams of sodium and jwtassium carbonate mix, 2:1, or 
sodium carbonate alone, in a platinum dish. (Prolonged fusion is not nec- 
essary.) The melt is cooled and then extracted with hot water to dissolve 
out the alkali sulphates. Barium carbonate, together with the other insoluble 
carbonates, may now be dissolved by hot dilute hydrochloric acid. From this 
solution barium may be precipitated by addition of sulphuric acid. If it is 
desired to separate barium along with strontium, calcium, and magnesium, 
the members of the preceding group)s are removed by US in acid and in ammo- 
niacal solution, as directed under *^ Separations." 

Sulphides. The ore is oxidized, as directed for pyrites under the subject 
of sulphur. After the removal of the soluble sulphates, the residue, containing 
silica, barimn, and small amounts of insoluble oxides, is fused and dissolved 
according to the procedure for sulphates. 

Carbonates. In absence of sulphates the material may be dissolved with 
hydrochloric acid, taken to dryness to dehydrate silica and after heating for 
an hour in the steam oven (110°d=) the residue is extracted with dilute hydro- 
chloric acid and filtered. The filtrate is examined for barium according to 
one of the procedures given later. 

Salts Soluble in Water. Nitrates, chlorides, acetates, etc., are dissolved 
with water shghtly acidulated with hydrochloric acid. 

Material Containing Organic Matter. The substance is roasted to destroy 
organic matter before treatment with acids or by fusion with the alkali carbonates. 

The Insoluble Residue remaining from the acid treatment of an ore may 
contain barium sulphate in addition to silica, etc. The filter containmg this 
residue is burned and the ash weighed. Silica is now volatilized by addition 
of hydrofluoric acid with a few drops of sulphuric acid, and evaporation to 
dryness. If an insoluble substance still remains after taking up the remaining 
residue with dilute hydrochloric acid, barium sulphate is indicated. This is 
treated according to the method given for sulphates. 

Note. The insoluble substance remaining is frequently ignited and weighed as 
barium sulphate without fusion with the carbonate. 

SEPARATIONS 

The Alkaline Earths 

Preliminary Considerations. In the determination of barium, calcium^ 
and strontium, the following causes may lead to loss of the elements sought: 

a. Presence of Phosphates, Phosphoric acid, free or combined, has a decided 
influence uix)n the determination of the members of this group. Combined 
as phosphate it will cause the complete precipitation of barium, calcium, and 
strontium, along with iron, almnina, etc., up>on making the solution alkaline 
for removal of the ammonium sulphide group. It is a common practice to 
hold up the iron-|-alumina by means of tartaric, citric, or other organic acids 
before making ammoniacal for precipitation of this group as oxalates, or again 
the basic acetate method is used for precipitation of iron and alumina; calcium, 



52 BARIUM 

barium, and strontium going into solution. These procedures may be satis- 
factory for the analysis of phosphate rock and similar products, but do not cop)e 
with the difficulty when large amounts of phosphates are present. In samples 
containing free phosphoric acid, barium, calcium, and strontium, present in 
small amounts, may remain in solution in presence of sulphates or oxalates » 
Appreciable amounts of calcium, 1% or more, may escap)e detection by the 
usual method of precipitation by ammonium oxalate added to the alkaline 
solution, on account of this interference, so that the removal of phosphoric acid 
before precipitation of this group is frequently necessary. This may be ac- 
complished by addition of potassium carbonate in sufficient excess to combine 
completely with the phosphoric acid and form carbonates with the bases. The 
material taken to dryness is fused with additional potassium carbonate in an 
iron crucible, and the fusion leached with hot water — sodium phosphate dis- 
solves and the carbonates of the heavy metals remain insoluble. 

b. Another source of loss is the presence of sulphates, either in the original 
material or by intentional or accidental addition, in the latter case due to the 
oxidation of hydrogen sulphide, which has been passed into the solution during 
the removal of elements of the hydrogen sulphide and ammonium sulphide 
groups, barium and strontium sulphate being precipitated along with these 
members. A potassium carbonate fusion will form NajS04, which may be 
leached out with water. 

c. Loss may be caused by occlusion of barium, calcium, strontium, and mag- 
nesium by the gelatinous precipitates Fe(OH)a, Al(OH)a, etc. A double precipita- 
tion of these compfcnds should be made if considerable amounts are present. 

d. A large excess of anmaonium salts, which accumulate during the pre- 
liminary separations, will prevent precipitation of the alkaline earths. This 
can be avoided by using the necessary care required for accurate work, the 
additioji of reagents by means of burettes or according to definite me^ure- 
ments in graduates, etc. Careless addition of large amounts of ammonium 
hydroxide and hydrochloric acid should be guarded against. In case large 
amounts of ammonium chloride are present, time is frequently saved by a 
repetition of the separations. Ammonium chloride may be expelled by heating 
the material, taken to dryness in a large platinum dish, the ammonium salts 
being volatilized. 

e. Carbon dioxide absorbed by ammonium hydroxide from the air will 
precipitate the alkaline earths with the ammonium sulphide group. 

Direct Precipitation on Original Sample. For the determination of 
barium, calcium, and strontium, it is advisable to take a fresh sample, rather 
than one that has been previously employed for the estimation of the hydrogen 
sulphide and ammonium sulphide groups, as is evident from the statements 
made above. The alkaline earths are isolated by being converted to the insol- 
uble sulphates and separations effected as given later under Sulphate Method. 

Preliminary Tests. Much time may be saved by making a preliminary 
test for barium, strontium, and calcium by means of the spectroscope and 
avoiding unnecessary separations. By this means one-thousandth of a 'milli- 
gram of barium, six hundred-thousandths of a milligram of strontium or calcium 
may be detected. The characteristic spectra of these elements are given in 
the chart. Plate II. 

By means of the spectroscope with the use of the ordinary Bunsen flame 
0.2 milligram of calcium, 0.6 milligram of strontium and 14 milligrams of barium 



# 



BARIUM 53 

may be detected per cc. The test is very much more delicate by the arc spectra 
method.^ The liquid containing the substance is connected to the positive 
pole and an iridium needle is connected by means of an adjustable resistance 
of 300 to 500 ohms to the negative pole. An E.M.F. of 100 to 200 volts and 
1 ampere current are required. By the arc it is possible to detect 0.002 milli- 
gram of calciimi, 0.003 milligram of strontium, 0.006 milligram of barium, 
0.1 milligram of magnesium per cc. In these concentratioas, calcium show's 
one brilliant line (423 nn)y a bright line (616 nn)j and a faint line between them; 
strontium two bright lines (422 and 461 mm) and two fairly bright lines; barium 
two brilliant lines (455 and 493 /i/*)» two other bright lines, and a fairly bright 
one; and magnesium a brilliant band composed of three linos (516.8 to 518.4 
/t/ji), as well as a fairly bright line fiulher towards the violet end of the spectrum. 

The flame test may be of value in absence of sodium; barium giving a green 
flame, strontimn a brilliant scarlet, and calcium an orange red. 

Separation from Members of Previous Groups. The members of the 
previous groups may be removed by precipitation as sulphides by H2S passed 
into the acid and then the alkaline solutions, the combined filtrates concen- 
trated to about 300 cc. and nmde slightly acid with hydrochloric acid. The fol- 
lowing procedures for isolation of barium from magnesium and the alkalies 
and from members of the alkaline earth group may be necessary before pre- 
cipitation in its final form. The methods of separation will apply to the analyses 
of the elements mentioned so that the details of procedure will not be given 
elsewhere. 

Separation of the Alkaline Earths from Magnesitmi and the Alkalies. 
Two general procedures will cover conditions commonly met with in analytical 
work: 

A. Oxalate Method. Applicable in presence of comparatively large 
portions of calcium. The acid solution containing not over 1 gram of the 
mixed oxides is brought to a volume of 350 cc. and for every 0.1 gram of mag- 
nesiimi present about 1 gram of ammonium chloride Ls added, unless already 
present. Sufiicient oxalic acid is added to completely precipitate the barium, 
calcium, and strontium.* (H2a04-2H20 = 126.04, 'Ba = 137.37, Ca =40.07, 
Sr =87.63.) The solution is slowly neutralized by addition, drop by drop, of 
dilute ammonium hydroxide (1 : 10), methyl orange being used as indicator. 
About i gram of oxalic acid is now added in excess, the solution again made 
alkaline with anunonium hydroxide, and allowed to settle for at least two hours. 
The precipitate is filtered off and washed with water containing 1% ammonium 
oxalate, faintly alkaline with ammonia. 

The precipitate contains all the calcium and practically all of the barium 
and strontium. If Mg is present in amounts of 10 to 15 times that of the 
alkaline earths a double precipitation is necessary, to remove it completely from 
this group. The oxalates are dissolved in hydrochloric acid and reprecipi- 
tated with ammonium oxalate in alkaline solution. 

The filtrate contains magnesium and the alkalies. Traces of bariiim and 
strontium may be present. If the sample contains a comparatively large 
proportion of barium and strontium, the filtrate is evaporated to drjTiess, 
the ammonium salts expelled by gentle ignition of the residue, and the Ba and 

IE. H.Riesenfeld and G. Pfutzer, Ber., 1913, 46, 3140-3144; Analyst, 1913, 38, 584. 
'Calcium and strontium will slowly precipitate in the oxalic acid solution. Ba 
oxalate will precipitate upon making the solution alkaline. 



54 BARIUM 

Sr recovered as sulphates according to the method described below. Mag- 
nesium is precipitated as magnesium ammonium phosphate from the filtrate. 

The oxalates of barium, calcium, and strontium are ignited to oxides, in which 
form they may be readily converted to chlorides by dissolving in hydrochloric 
acid, or to nitrates by nitric acid. 

B. Sulphate Method. Applicable in presence of comparatively large pro- 
portions of barium, strontium, or magnesium. The solution containing the 
alkaline earths, magnesium and the alkalies is evaporated to dryness and 
about 5 cc. concentrated sulphuric acid added, followed by 50 cc. of 95% 
alcohol. The sulphates * of barium, calcium, and strontium, are allowed to 
settle, and then filtered onto an S. and S. No. 589 ashless filter paper and washed 
with alcohol until free of magnesium sulphate. In presence of large amounts 
of magnesium as in case of analyses of Epsom salts and other magnesium salts 
it will be necessary to extract the precipitate by adding a small amount of water, 
then sufficient 95% alcohol to make the solution contain 50% alcohol and 
filter from the residue. Magnesium is determined in the filtrate. 

The residue containing barium, calcium, and strontium as sulphate is 
fused with 10 parts of potassium carbonate or sodium acid carbonate until the 
fusion becomes a clear molten mass, a deep platinum crucible being used for 
the fusion. A platinum wire is inserted and the mass allowed to solidify. The 
fusion may be removed by again heating until it begins to melt around the 
surface next to the crucible, when it may be lifted out on the wire. The mass 
is extracted with hot water and filtered, Na«S04 going into the solution and 
the carbonates of barium, strontium, and calcium remaining insoluble. The 
carbonates should dissolve completely in hydrochloric acid or nitric acid, other- 
wise the decomposition has not been complete, and a second fusion of this 
insoluble residue will be necessary. 

Separation of the Alkaline Earths from One Another. This separation 
may be effected by either of the following processes: 

1. Barium is separated in acetic acid solution as a chromate from strontium 
and calcium; strontiiun is separated as a nitrate * from calcium in ether-alcohol 
or amyl alcohol. 

2. The three nitrates are treated with ether-alcohol in which barium and 
strontium nitrates are insoluble and calcium dissolves; the barium is now 
separated from strontium by ammonium chromate. 

Procedures. 1. (a) Separation of Barium from Strontitmi (and from 
Calcitmi). In presence of an excess of ammonium chromate, barium is pre- 
cipitated from solutions, slightly acid with acetic acid, as barium chromate 
(appreciably soluble in free acetic acid), whereas strontium and calcium remain 
in solution. 

The mixed oxides or carbonates are dissolved in the least amount of dilute 
hydrochloric acid and the excess of acid expelled by evaporation to near drj^ness. 
The residue is taken up in about 300 cc. of water and 5-6 drops of acetic acid 
(sp.gr. 1.065) together with sufficient anunonium acetate (30% solution) to 
neutralize any free mineral acid present. The solution is heated and an excess 
of ammoniiun chromate (10% neutral sol.) ' added (10 cc. usually sufficient). 

^Solubility of BaSO4=0.17 milligram, Ca804 = 179 milligram, SrS04 = 11.4 milH- 
grams |)er 100 cc. sol. cold. 

» Method of Stromayer and Rose. H. Rose, Pogg. Ann., 110, 292, (1860). 

'The solution is prepared by adding NH4OH to a solution of (NH4)2Cr207 until 
yellow. The solution should be* left acid rather than alkaline. 



BARIUM 55 

The precipitate of barium chromate is allowed to settle for an hour and filtered 
off on a small filter and washed with water containing ammonium chromate 
until free oi soluble strontium and calcium (test — ^addition of NH4OH and 
(NH4)2COi produces no cloudiness), and then with water until practically free of 
ammonium chromate (e.g., only slight reddish brown color with silver nitrate 
solution). 

To separate any occluded precipitate of strontium or calcium the filter 
paper is pierced and the precipitate rinsed into a beaker with warm dilute nitric 
acid (sp.gr. 1.20) (2 cc. usually are sufficient). The solution is diluted to about 
200 cc. and boiled. About 5 cc. of ammonium acetate, or enough to neutralize 
the free HNOa, are added to the hot solution and then sufficient ammonium 
chromate to neutralize the free acetic acid, 10 cc. usually sufficient. The washing, 
as above indicated, is repeated. Barium is completely precipitated and may 
be determined either as a chromate or a sulphate or by a volumetric pro- 
cedure. Strontium and calcium are in the filtrates and may be separated as 
follows: 

(b) Separation of Strontium from Calcium. The method depends upon 
the insolubility of strontium nitrate and the solubility of calcium nitrate in 
a mixture of ether-alcohol, 1:1. 

Solubility of SrNO, = l part SrNO, m 60,000 parts of the mbcture. Ca 
easily soluble. 

If the solution is a filtrate from barium, 1 cc. of nitric acid is added and 
the solution heated and made alkaline with ammonium hydroxide followed 
immediately with ammonium carbonate, the carbonates of strontium (together 
with some SrCr04) and calcium will precipitate. The precipitate is dissolved 
in hydrochloric acid and reprecipitated from a hot solution with ammonium 
hydroxide and ammonium carbonate. The precipitate,. SrCOa and CaCOa, 
is washe^l once with hot water and is then dissolved in the least amount of 
nitric acid, washed into a small casserole, evaporated to dryness and heated 
for an hour at 140 to 160° C. in an oven, or at llO** C. over night. The dry 
mass is pulverized and mixed with 10 cc. of ether-alcohol (absolute alcohol, one 
part, ether-anhydrous, one part). Several extractions are thus made, the extracts 
being decanted ofif into a flask. The residue is again dried in an oven at 140 
to 160° O., then pulverized and washed into the flask with the ether-alcohol 
mixture and digested for several hours with frequent shaking of the flask. The 
residue is washed onto a filter moistened with ether-alcohol mixture. Strontium 
nitrate, Sr(N08)2, remains insoluble, and may be dissolved in water and de- 
termined gravimetrically as a sulphate, oxide, or carbonate or volume trically. 
Calcium is in the filtrate and may be determined gravimetrically as an oxide 
or volumetrically. 

Instead of using a mixture of ether-alcohol, amyl alcohol may be used (hood) , 
the mixture being kept at boiling temperature to dehydrate the alcohol to pre- 
vent solution of strontium (6.p. =130° C.). 

2. Separation of Baritmi and Strontium from Calcium.^ The procedure 
depends upon the insolubility of barium nitrate, (BaN0j)2, and strontium nitrate, 
Sr(N0a)2, in a mixture of anhydrous ether and absolute alcohol or anhydrous 
amyl alcohol, whereas Ca(N03)2 dissolves. 

The mixed oxides or carbonates are dissolved in nitric acid and taken to 
dryness in a beaker or Erlenmeyer flask, and heated for an hoiur or more in an 

»See Fresenius, Z. anal. Chem., 29, 413-430 (1890). 



56 BARIUM 

oven at 140 to 160° C. Upon cooling, the mixture is treated with ten times 
its weight of ether-alcohol mixture and digested, cold, in the covered beaker 
or corked flask for about two hours with frequent stirring. An equal volume 
of ether is now added and the digestion continued for several hours longer. 
The residue is washed by decantation with ether and alcohol mixture until 
calcium is removed (test — no residue on platinum foil with drop of filtrate evap- 
orated to dr>Tiess). If calcium is present in amount above 0.5 gram, the residue 
is dissolved in a little water, again evaporated and dried and then extracted 
with ether-alcohol as directed above. 

Calcium is in the filtrate and may be determined by precipitation as a sul- 
phate in the alcohol solution or as an oxide by evaporation of the ether-alcohol 
and precipitation as calcium oxalate, CaCjOi, according to directions given in 
the determination of calcium. 

Barium and strontium may be separated by precipitation of barium as a 
chromate, the nitrate residue being dissolved in water and barium precipitated 
according to directions given under Procedure No. 1. 

Amyl alcohol may be used in place of ether-alcohol by digesting the nitrates 
in a boiling solution (130° C), calcium going into solution and barium and 
strontium remaining insoluble as nitrates. 



GRAVIMETRIC METHODS FOR THE DETERMINATION OF 

BARIUM 

For reasons given under *' Preliminary Considerations," it is advisable to 
take a special sample for the determination of barium that has not undergone 
treatment with hydrogen sulphide or ammonium hydroxide, since these may 
cause the loss of barium as stated. 

Preparation of the Sample. The following general schemes will meet 
practically all conditions: 

Barium in Insoluble Residue. In the complete analysis of ores the residue 
remaining insoluble in acids is composed largely of silica, together with difficultly 
soluble substances, among which is barium sulphate. This residue is best fused in a 
platinum dish with sodium carbonate or a mixture of sodium and potassium car- 
bonates (long fusion is not necessary). The cooled mass is digested with hot water 
to remove the soluble sodium compounds, silicate being included. Barium, to- 
gether with the heavy metals, remains insoluble as carbonate and may be filtered 
off. The residue is now treated with dilute ammonia water to remove the 
adhering sulphates (testing the filtrate with hydrochloric acid and barium chlo- 
ride solution; the washing being complete when no white precipitate of barium 
sulphate forms). The carbonates are washed off the filter into a 500-cc. beaker, 
the clinging carbonate being dissolved by pouring a few cc. of dilute, 1 : 1, hydro- 
chloric acid on the pai^er placed in the funnel. This extract is added to the pre- 
cipitate in the beaker and the latter covered to prevent loss by spattering. 
Additional hydrochloric acid is cautiously added so that the precipitate com- 
pletely dissolves and the solution contains about 10 cc. of free hydrochloric 
acid (sp.gr. 1.2). Barium is precipitated from this solution best as a sulphate 
according to directions given later. 

Silicates. One gram of the finely pulverized sample is treated with 10 



BARIUM 57 

cc. of dilute sulphuric acid, 1 : 4, and 5 cc. of strong hydrofluoric acid. The 
mixture, evaporated to small bulk on the steam bath, is taken to SOs fumes 
on the hot plate. Additional sulphuric acid and hydrofluoric acid are used if 
required. By this treatment the silica is expelled and barium, together with 
other insoluble sulphates, will remain upon the filter when the residue is treated 
with water and filtered. Lead sulphate, if present, may be removed by washing 
the residue with a solution of ammonium chloride. Barium sulphate may be 
purified by fusion with potassium carbonate as above directed or by dissolving 
in hot concentrated sulphuric acid, and precipitating again as BaSOi by dilution. 
Ores may be decomposed by either of the above methods or a combination 
of the two. Sulphide ores require roasting to oxidize the sulphide to sulphate. 

Determination of Barium as a Chromate 

A preliminary spectroscopic test has indicated whether a separation from 
calcium and strontium is necessary. If these are present, barium is separated 
along with strontium from calcium as the nitrate in presence of alcohol-ether 
mixture, according to directions given under " Separations. '* Barium is now 
precipitated as the chromate, BaCr04, from a neutral or slightly acetic acid 
solution, strontium remaining in solution. 

Precipitation of Barium Chromate. If barium is present in the form 
of nitrate, together with strontium, the mixed nitrates are evaporated to dryness 
and then taken up with water. About 10 cc. ammonium acetate (300 grams 
NH4C,H,0, neutralized with NH4OH+H2O to make up to 1000 cc.) added 
and the solution heated to boiling. Five cc. of 20% ammonium bichromate 
are added drop by drop with constant stirring and the precipitate allowed t3 
settle until cold. The solution is decanted off from the precipitate through 
a filter and washed by decantation with dilute (0.5%) solution of ammonium 
acetate, until the excess chromate is removed, as indicated by the filtrate passing 
through uncolored. If much strontium was originally present, a double pre- 
cipitation is necessary, otherwise the precipitate may be filtered directly into 
a Gooch crucible and ignited, the following paragraph directions being omitted. 

Purificadoii from Strontium. The precipitate is dissolved from the filter 
by running through dilute (1:5) warm nitric acid, poured upon the chromate, 
catching the solution in the beaker in which the precipitation was made; the 
least amount of acid necessary to accomplish this being used and the filter 
washed with a little warm water. Ammonium hydroxide is now added to the 
solution, cautiously, until a slight permanent precipitate forms and then 10 cc. 
of ammonium acetate solution added with constant stirring and the mixture 
heated to boiling. The precipitate is allowed to settle until the solution is cold 
and then filtered and washed by decantation as before, a Gooch crucible being 
used to catch the precipitate. 

Ignition. The precipitate is washed once with dilute alcohol, 1 : 10, dried 
at 110** C, and ignited, gently at first and then to a dull red heat until the 
color of the chromate is uniform. It is advisable to cover the crucible at first 
and then after five minutes to remove the cover. 

BaCr04X 0.6051 =BaO. BaCr04X 0.5420 =Ba. 

Notes. The use of sodium hydrate or acetate in place of the ammonium 
hydroxide and acetate is sometimes recommended, owing to the slight solubility of 



58 BARIUM 

BaOOi in ammonium salts, as seen by the following table, approximate figures being 

given: 

100,000 parts of cold water dissolves 0.38 parts BaCrOi 

100,000 parts of hot water dissolves 4.35 parts BaCr04 

100,000 of 0.5% solution of NH4CI dissolves 4.35 parts BaCrOi 

100,000 of 0.5% solution of NH4NO8 dissolves 2.22 paits BaCrO* 

100,000 of 0.75% solution of NH4C,HaO, dissolves 2.00 parts BaCrO* 
100,000 of 1.5% solution of NH4C2H,02 dissolves 4.12 parts BaCr04 
100,000 of 1% acetic acid dissolves 20.73 parts BaCr04 

Although the solvent action of ammonium salts is practically negligible under 
conditions of analysis given above, the solvent action of free acetic acid is of importance, 
so that it ia necessary to neutralize or eliminate free mineral acids before addition 
of the acetate salt. 

The edges of the BaCr04 piecipitate upon drying may appear green^ owing to 
the action of alcohol; upon ignition, however, the yellow chromate is obtamed. The 
color orange yellow, when hot, fades to a light canary yellow upon cooling. 

BaCr04, molwt., 253.47; sp.gr., 4.498^*°; 100 cc. H2O sol. cold will dissolve 
0.00038"'' gram, hot dissolves 0.0043 gram; soluble in HCl, HNO^, yellow rhombic 
plates. 

Determination of Barium by Precipitation as Sulphate, BaS04 

This method depends upon the insolubility of barium sulphate in water 
and in very dilute hydrochloric acid or sulphuric acid, one gram of the salt 
requiring about 344,000 cc. of hot water to effect solution. 

Reaction^ BaCl,+H2S04 = BaS04+2HCl. 

BaS04, mol.wLj 233,44; sp.gr,^ 4.47 and 4.33; m.p.f 1580** {amorphous decom- 
poses) ; H2O dissolves 0.0001 72°° gram and 0.0003"*" per 100 cc, 3% RCldissolves 
0.0036 gram. Soluble in cone. H2SO4. White, rhombic and amorphous forms. 

Procedure. The slightly hydrochloric acid solution of barium chloride, 
prepared according to directions given, is heated to boiling (volume about 200- 
300 cc.) and a slight excess of dilute hot sulphuric acid added. The precipitate 
is settled on the water bath and the clear solution then decanted through a 
weighed Gooch crucible or through an ashless filter paper (S. and S. 590 quality). 
The precipitate is transferred to the Gooch (or paper), and washed twice with 
very dilute sulphuric acid solution (0.5% H2SO4), and finally with hot water 
until free of acid. The precipitate is dried and ignited, at first gently and then 
over a good flame to a cherry red heat, for half an hour. The residue is weighed 
as barium sulphate, BaS04. 

BaS04 X 0.5884 =Ba, or X0.6569=BaO, or X 0.8455 =BaCO,. 

N0TE8. The determination of barium is the reciprocal of the determination of 
sulphur or sulphuric acid. Precautions and directions given for the sulphiu: pre- 
cipitation apply here also, with the exception that dilute sulphuric acid is used, as 
the precipitating reagent in place of barium chloride. 

The author found that precipitation of barium sulphate in a large volume of cold 
solution containing 10 cc. of concentrated hydrochloric acid per 1600 cc. of solution, 
by adding a slight excess of cold dilute sulphuric acid in a fine stream, exactly in 
the manner that barium chloride solution is added in the j)reeipitation of sulphur, 
and allowing the precipitate to settle, at room temperature, for several hours (pref- 
erably over night), gives a precipitate that is pure and does not pass through the 
Gooch asbestos mat. We refer to the chapter on Sulphur for directions for filtering, 
washing, and ignition of the residue. 



BARIUM 69 

VOLUMETRIC METHODS FOR THE DETERMINATION OF 

BARIUM 

Titration of the Barium Salt with Dichromate 

This method is of value for an approximation of the amount of barium 
present in a solution that may also contain calcium, strontium, and magnesium 
or the alkalies. It depends upon the reaction 

2BaCl2+K2Cr207+H20=2BaCr04+2KCl+2HCl. 

N/10 K2Crt07 (precipitation purposes) contains 7.355 grams pure salt per 
liter. 

Procedure. The solution containing the barium is treated with ammonia 
until it just smells of it. (If an excess of ammonia is present the solution is made 
faintly acid with acetic acid.) It is then heated to about 70** C. and the 
standard dichromate added, with stirring until all the barium is precipitated 
and vthe clear supernatant solution is a faint yellow color from the slight 
excess of the reagent. For accurate work it is advisable to titrate the pre- 
cipitate formed by one of the methods given below. One cc. K2Cr207 =0.00687 
gram Ba. (Note reaction given above.) 

Note. An excess of potassium dichromate maybe added, the precipitate filtered 
off, washed and the excess of dichromate determined as stated below. 

Reduction of the Chromate with Ferrous Salt and Titration with 

Permanganate 

Ferrous sulphate reacts with barium chromate as follows: 

2BaCr04+6FeSOt+8H2S04=3Fe,(S04)3+Cr2(S04)a+2BaS04+8H,0. 

An excess of ferrous salt solution is added and the excess determined by 
titration with N/10 KMn04 solution. Fe = JBa. 

Reagents. N/10 solution of KMn04. N/10 FeS04 (27.81 grams per liter) 
or FeSO* • (NH4),S04 (39.226 grams per liter). One cc. =0.004579 Ba. 

Procedure. The well-washed precipitate of barium chromate is dissolved 
in an excess of standard N/10 ferrous ammonium sulphate solution containing 
free sulphuric acid. The excess ferrous salt is titrated with standard N/10 
potassium permanganate solution. 

(Cc. N/10 ferrous solution minus cc. permanganate titration) multiplied by 
0.004579 gives grams barium in the solution. Iron factor to barium is 0.8187. 

Potassium Iodide Method 

The procedure depends upon the reactions: 

1. 2BaCr04+6KH-16HCl =2BaCl2+2CrCla+6KCH-8H20+6I. 

2. 3I,+6Na,S,0,=6NaI+3Na2S40e. 

Procedure. The precipitate, BaCrOi, is dissolved in 50 to 100 cc. of 
dilute hydrochloric acid and about 2 grams of solid potassium iodide salt added 
and allowed to react about ten minutes. The liberated iodine is now titrated 



60 BARIUM 

with N/10 thiosulphate. Near the end of the titration starch solution is added 
and followed by N/10 thiosulphate until the color disappears. 

One cc. N/10 NajSjO, =0.004579 gram Ba. 

Titration of Barium Carbonate with Standard Acid 

To the well-washed barium carbonate, BaCOa, an excess N/10 H2SO4 is added 
and the excess acid determined. 

One cc. N/10 acid =0.00687 gram Ba. 

ANALYSIS OF BARYTES AND WITHERITE 

Barytes or heavy spar is a variety of native barium sulphate, and witherite 
a native barium carbonate. These minerals are typical examples of barium- 
bearing ores. The analysis may involve the determination of barium and 
calcium sulphates or carbonates, magnesia, iron and aluminum oxides and 
moisture. Traces of lead, copper, and zinc may be present, as well as sulphide, 
sulphur and fluorine in fluorspar. The following is an approximate composition 
of a high-grade sample: 

BaS04=96%, CaC03 = 1.5%, MgCO,=0.3%, SiO,=0.5%, Al,O,=0.5%, 

Fe^Oa =0.2%, H2O =0.5%. 

For completo analysis treat as directed under preparation of the sample. 

Procedure for Commercial Valuation of the Ore 

Barium Sulphate and Silica 

One gram of the finely pulverized sample is digested with about 50 cc. of 
concentrated hydrochloric acid and taken to dryness on the steam bath. The 
residue is taken up with 50 cc. of water, 10 cc. of hydrochloric acid added, 
and the mixture heated on the steam bath for ten minutes, then heated to boil- 
ing and filtered. The residue of barium sulphate and silica is washed well with 
hot water containing a little hydrochloric acid and finally with pure water. It 
is now ignited and weighed as BaS04+Si02, or total insoluble matter. 

The residue in a platinum dish is now treated with a little hydrofluoric 
acid + sulphuric acid, and silica expelled as usual. The residue ignited ^BaSO*. 

Silica = difference between total insoluble matter and BaSOi. 

Barium Carbonate 

Barium, originally present as a carbonate, will be found in the filtrate together 
with iron, alumina, etc., and may be precipitated l)y addition of sulphuric acid. 
Barium sulphate is filtered off, washed, ignited, and weighed. BaS04X 0.84555 
=BaCO,. 

Iron and Alumina Oxides 
These are determined in the filtrate from barium precipitation in the usual way. 



BARIUM 61 

Calcium and Magnesium 

Detennined in the filtrate from iron and alumina by the regular procedures. 

Soluble Sulphates 

One gram of the powdered sample is boiled with 20 cc. cone. HCl and 200 
cc. water, the insoluble residue filtered ofif and washed. The filtrate contains 
the soluble sulphate. This may be precipitated by addition of BaCU solution 
according to the procedure for sulphur. B aSOiX 0.5833 ^CaSOi. BaS04X 
0.2402 =CaO. 

If lime, CaO, thus calculated, is less than lime precipitated as oxalate, the 
difference is calculated to CaCOa if CO2 is present, otherwise to CaO. 

Loss on Ignition 
Represents water free and combined, carbon dioxide and organic matter. 



BISMUTH 

Wilfred W. Scott 

BUai.wt. 208.0; 8p.gr. 9.7474; m.p. 271'';^ b.p. 14!S0° C; oxides, 

Bi20„ BisOft. 

DETECTION 

Bismuth is precipitated from its solution, containing free acid, by HsS gas, 
as a brown sulphide, Bi28s. The compound is insoluble in ammonium sulphide 
(separation from arsenic, antimony, and tin), but dissolves in hot dilute nitric 
acid (separation from mercury). The nitrate, treated with sulphuric acid and 
taken to SOj fumes, is converted to the sulphate and dissolves upon dilution 
with water (lead remains insoluble as PbSOi). Bismuth is precipitated from 
this solution by addition of ammonium hydroxide, white Bi(OH)j being formed 
(copper and cadmium dissolve). If this hydroxide is dissolved with hydro- 
chloric acid and then diluted with a large volume of water, the white, basic 
salt of bismuth oxychloride, BiOCl, is precipitated. The compound dissolves 
if sufficient hydrochloric acid is present. It is insoluble in tartaric acid (dis- 
tinction from antimony). 

Reducing Agents. Formaldehyde in alkaline solution, hypophosphorous 
acid, potassium or sodium stannite, reduce bismuth compounds to the metallic 
state. For example, a hot solution of sodium stannite poured onto the white 
precipitate of Bi(OH)j on the filter will give a black stain. The test is very 
delicate and enables the detection of small amounts of the compound. 

3K,SnO,+2BiCl3+6KOH=2Bi+3K,SnO,+6KCl-|-3H,0. 

Blowpipe Test. A compound of bisnmth heated on charcoal with a 
powdered mixture of carbon, potassium iodide and sulphur, will give a scarlet 
incrustation on the charcoal. 

ESTIMATION 

The determination of bismuth is required in complete analysis of ores of 
cobalt, nickel, copper, silver, lead, and tin, in which it is generally found 
in small quantities. In evaluation of bismuthite, bismuth ochre, etc. In the 
analysis of the minerals wolfram, molybdenite. It is determined in the residues 
from the refining of lead (the principal source of bismuth in the United 
States). In the analysis of alloys — antifriction metals, electric fuses, solders, 
stereotype metals, certain amalgams used for silvering mirrors (with or with- 
out lead or tin), and in bismuth compounds. 

Preparation and Solution of the Sample 

In dissolving the substance, the following facts must be kept in mind: nitric 
acid is the best solvent of the metal. Although it is soluble in hot sulphuric 
acid, it is only very slightly soluble in the cold acid. The metal is practically 
insoluble in hydrochloric acid, but readily dissolves in nitrohydrochloric acid. 

* U. S. Bureau of Standards Cir. 35. 

62 



BISMUTH 63 

The hydroxides, oxides, and most of the bismuth salts are readily soluble in 
hydrochloric, nitric, and sulphuric acids. 

Ores or Cinders. One gram of the finely pulverized ore or cinder (or larger 
amounts where the bismuth content is very low) is treated in a 4(X)-cc. beaker 
with 5 cc. of bromine solution (Br+KBr+HjO)/ followed by the cautious 
addition of about 15 cc. of HNOi (sp.gr. 1.42). When the violent action has 
ceased, which is apt to occur in sulphide ores, the mixture is taken to dryness 
on the steam bath, 10 cc. of strong HCl and 20 cc. of concentrated H2SO4 and 
the covered sample heated until SOj fumes are freely evolved. The cooled solu- 
tion is diluted with 50 cc. of water and gently heated imtil only a white or 
light gray residue remains. The solution is filtered and the residue washed 
with dilute HjSOi (1 : 10), to remove any adhering bismuth. Silica, the greater 
part of the lead (also BaS04) remain in the residue, whereas the bismuth, to- 
gether with iron, alumina, copper, antimony, etc., are in the solution. Details 
of further treatment of the solution to effect a separation of bismuth are 
given under "Separations" and the procedures for determination of bismuth. 

Alloys, Bearing Metal, etc. One gram of the borings, placed in a small 
beaker, is dissolved by adding 20 cc. of concentrated HCl and 5 cc of strong 
HNOa. The alloy will usually dissolve in the cold, unless considerable lead 
is present, in which case prolonged heating on the steam bath may be neces- 
sary. (A yellow or greenish-yellow color at this stage indicates the presence of 
copper.) Lead may now be removed either as a sulphate by taking to SOj fumes 
with H2SO4 or by precipitating as a chloride, in the presence of alcohol, accord- 
ing to directions given under Separations. The bismuth is determined in the 
filtrate from lead according to one of the procediu-es given under the quanti- 
tative methods. 

Lead Bullion, Refined Lead.* Ten to twenty-five grams of the lead, 
hammered or rolled out into thin sheets and cut into small pieces, are taken 
for analysis. The sample is dissolved by a mixture of 250 cc. of water and 
40 cc. of strong nitric acid, in a large covered beaker, by warming gently, pref- 
erably on the steam bath. When the lead has dissolved, the beaker is removed 
from the heat and dilute ammonia (1:2) added to the warm solution, very 
cautiously and finally drop by drop until the free acid is neutralized and the 
liquid remains faintly opalescent, but with no visible precipitate. Now 1 cc. 
of dilute HCl (1 : 3) is added. The solution will clear for an instant and then 
a crystalline precipitate of bismuth oxychloride will form, if any considerable 
amount of bismuth is present. The beaker is now placed on the steam bath 
for an hour, during which time the bismuth oxychloride will separate out, 
together with a small amount of lead and with antimony if present in appre- 
ciable amounts. The further isolation and purification of bismuth is given 
imder "Separations." In brief — antimony is removed by dissolving the pre- 
cipitate in a small amount of hot dilute HCl (1 : 3), precipitating bismuth, 
traces of lead, and the antimony by H2S, dissolving out the antimony sulphide 
with warm ammonium sulphide, dissolving the BiSz and PbS in HNO3 and 
reprecipitation of the bismuth according to the procedure given above. Bis- 
muth is now determined as the oxychloride. Further details of this method 
are given under the gravimetric procedures for bismuth. 

* Bromine solution is made by dissolving in water 75 grams of KBr, to which are 
added 50 grams of liquid bromine and the mixture diluted to 500 cc. with water. 

* Bismuth in Refined Lead. "Technical Methods of Ore Analysis." A. H. Low. 



64 BISMUTH 



SEPARATIONS 

T?ie foUowing procedures are given in the order thai would he followed in tht 
complete analysis of an ore^ in v:hich all the canstituents are sought. This general 
scheme^ however, is not required for the majority of bismuth-bearing samples com- 
monly met with in the commercial laboratory , direct precipitations of bismuth fre- 
quently being possible. 

Separation of Bismuth from Members of Subsequent Groups, Fe, Cr, 
Al, Mn, Co, Ni, Zn, Mg, the Alkaline Earths and Alkalies, together with 
Rare Elements of these Groups. The solution should contain 5 to 7 cc. of 
concentrated hydrochloric acid (sp.gr. 1.19) for every 100 cc. of the sample. 
The elements of the hydrogen sulphide group are precipitated by saturating 
the solution with H2S (Hg, Pb, Bi, Cu, Cd, As, Sb, Sn, Mo, Se, Te, Au, Pt). 
The members of subsequent groups remain in solution and pass into the filtrate. 

Separation of Bismuth from Arsenic, Antimony, Tin, Molybdenum, 
Tellurium, Selenium. In presence of mercury, the soluble members of the 
hydrogen sulphide group are separated from the insoluble sulphides by digest- 
ing the precipitate above obtained with ammonium sulpliide; in absence of 
mercury, however, which is generally the case, digestion of the sulphides with 
sodium hydroxide and sodium sulphide solution is preferred, the general pro- 
cedure being followed. Mercurj", lead, bismuth, copper, and cadmium remain 
in the residue, whereas the other members of the group dissolve. 

Separation of Bismuth from Mercury. The insoluble sulphides, remain- 
ing from the above treatment with ammonium sulphide after being washed 
free of the soluble members of this group, are placed in a porcelain dish and 
boiled with dilute nitric acid (sp.gr. 1.2 to 1.3). The solution thus obtained 
is filtered, upon dilution, from- the insoluble sulphide of mercury. A little 
of the lead may remain as PbSOi, the solution may contain lead, bismuth, copper, 
and cadmium. 

Separation of Bismuth from Lead. This is the most important pro- 
cediu^ in the determination of bismuth as the separation is almost invariably 
necessary, as these elements commonly occur together. Bismuth produced 
in the United States in 1912 was obtained entirely from the residues in the re- 
finmg of lead.^ 

There are two general procedures for the separation of lead and bismuth. 

A. Precipitation of lead either as lead sulphate or as lead chloride, the 
bisnmth remaining in solution under the conditions of the precipitation. 

B. Precipitation of bismuth as the oxy chloride or subnitrate, lead remaining 
in solution. 

Precipitating Lead as PbS04. This procedure is generally used in the 
process of a complete analysis of an ore containing lead and l)ismuth. The nitric 
acid solution of the sulphides, obtained upon removal of the soluble group and 
mercury by boiling the insoluble sulphides with dilute nitric acid, is treated 
with about 10 cc. of strong sulphuric acid, and taken to SOj fumes by heating. 
The cooled sulphate solution is diluted with water and the insoluble lead sul- 
phate filtered off and washed with dilute sulphuric acid solution (1 : 20), 
Bismuth passes into solution, together with copper and cadmium, if also present 
in the original sample. 

^ Mineral Industiy, 1912, p. 98. 



BISMUTH 65 

Predpitatioii of Lead as PbCli. This separation is used in the complete 
analysis of pig lead, the details of the separation being given under this 
subject. 

As the separation of bismuth from lead by precipitation of the former 
element as the oxychloride or subnitrate is incorporated in the quantitative 
methods following, it will not be taken up here. 

Separation of Bismuth from Copper and Cadmium. This separation is 
accomplished by precipitating bismuth as the oxychloride with hydrochloric acid, 
or as the carbonate by adding an excess of ammonium carbonate to the solu- 
tion nearly neutralized by ammonia, or as the hydroxide by adding an excess 
of ammonia. Details of these procedures are given under the gravimetric methods 
for determining bismuth. 



GRAVIMETRIC METHODS FOR THE DETERMINATION OF 

BISMUTH 

Determining Bismuth by Precipitation and Weighing as the 

Basic Chloride, BiOCl 

The determination depends upon the formation of the insoluble oxychloride, 
BiOCl, when a hydrochloric acid solution of bismuth is sufficiently diluted 
with water, the following reaction taking place, BiCU+HjO =BiOCl+2HCl. 

The procedure is recommended for the determination of bismuth in refined 
lead, bearing metal, and bismuth alloys. Copper, cadmium, and lead do not 
interfere; appreciable amounts of antimony and tin, however, should be re- 
moved by H^ precipitation and subsequent treatment with NajS, and the resid- 
ual sulphides dissolved in hot dilute nitric acid, according to directions given 
under "Separations." 

Properties of BiOCl. Mol.wt.j 259A6; sp.^., 7.717^*'; m.p., red heat; i/iso/. 
in HiO and in H2C4H40*, soluble in adds. Appearance is white j quadratic crys- 
talline form. 

Procedure. The solution of bismuth, freed from appreciable amounts of 
tin and antimony, is warmed gently and treated with sufficient ammonia to 
neutralize the greater part of the free acid. At this stage a precipitate is formed 
by the addition, which dissolves with difficulty; the last portion of the dilute 
ammonia (1 : 2) is added drop by drop, the solution is diluted to about 300 
cc, and the remainder of the free acid neutralized with dilute ammonia added 
cautiously imtil a faint opalescence appears, but not enough to form an appre- 
ciable precipitate. One to 3 cc. of dilute hydrochloric acid (1 part HCl 
sp.gr. 1.19 to 3 parts HjO) are now added, the mixture stirred and the bismuth 
oxychloride allowed to settle for an hour or so on the steam bath, then filtered 
hot by decanting ofif the clear solution through a weighed Gooch crucible. The 
precipitate is washed by decantation twice with hot water and finally washed 
into the Gooch, then dried at 100** C. and weighed as BiOCl. 

BiOCl X 0.8017 =Bi. 

Note. Three cc. of dilute hydrochloric acid (or 1 cc. cone. HCl, sp.gr. 1.19) 
are si^cicnt to completely precipitate 1 gram of bismuth from solution. 



66 BISMUTH 



Determination of Bismuth as the Oxide, Bi203 

Preliminary Considerations. The determination of bismuth as the oxide 
requires the absence of hydrochloric acid or sulphuric acid from the solution 
of the element, since either of these acids invariably contaminates the final 
product. In presence of these acids, which is frequently the case, determination 
of bismuth by precipitation as BiSi or by reduction to the metal and so weighing 
is generally recommended; a brief outline of the methods is given later; 
a solution of bismuth free from hydrochloric acid and practically free of sul- 
phuric acid may be obtained by precipitating BiaSx, together with CuS, CdS, 
and PbS, the amount of sulphuric acid formed by the reaction being negligible. 
Bismuth should be in a nitric acid solution, free from antimony and tin. 

Two general conditions will be considered : 1 . Solutions containing lead. 
Copper and cadmiiun may also be present. 2. Solutions free from lead. Copper 
and cadmium may be present. 

1. Separation from Lead, Copper, and Cadmium, by Precipitation as 
Basic Nitrate. ^ Either the sulphuric or hydrochloric acid methods may be employed 
for effecting the separation of lead by precipitation. Furthermore advantage may 
be taken of the fact that bismuth nitrate is changed by the action of water into an 
insoluble basic salt, while lead, copper and cadmium do not undergo such a trans- 
formation. 

Procedure. The bismuth nitrate solution is evaporated to syrupy con- 
sistency and hot water added with constant stirring with a glass rod. The 
solution is again evaporated to dryness, and the hot- water treatment repeated. 
Four such evaporations are genemlly sufficient to convert the bismuth nitrate 
completely into the basic salt; when this stage is reached the addition of water 
will fail to produce a turbidity. The solution is finally evaporated to dryness 
and, when free from nitric acid, is extracted with cold anmionium nitrate solu- 
tion (I.NH4NOJ : 500 HjO) to dissolve out the lead and other impurities. After 
allowing to stand some time with frequent stirring, the solution is filtered and 
the residue washed with anunonium nitrate solution, then dried. 

Ignition to Bismuth Oxide. As much of the precipitate as possible is 
transferred to a weighed porcelain crucible, the filter is burned and the ash 
added to the main precipitate. This is now gently ignited over a Bunsen 
burner. Too high heating will cause the oxide to fuse and attack the glaze 
of the crucible. 

Properties. Bi(0H),X03 mol.wi., 304.03; sp.gr., 4.928"**; decomp., 260°; 
insoL in HjO ; sol, in adds; hexagonal plates, 

BijOi =mo/. wt.f 464.0; sp.gr. , 8.8 to 9.0; w.p., 820 to 860°; insoluble in cold 
water and in alkalies, but soluble in acids; yellow tetragonal crystals. 

BijOa X 0.8965 =Bi. 

2. Precipitation of Bismuth as the Subcarbonate or Hydroxide, Lead 
being Absent. Either of these procedures effects a separation of bismuth from 
copper and cadmium. 

A. Procedure. Precipitation of the Subcarbonate. The solution is 
diluted to about 300 cc. and dilute anmionia added cautiously until a faint 
turbidity is obtained and then an excess of anmionium carbonate. The solution 

* J. Lowe, Jour. prak. Chem., (1), 74, 344, 1858. 



BISMUTH 



6/ 



is heated to boiling, the precipitate filtered off, washed with hot water, dried 
and ignited according to directions given in the bismuth subnitrate method. 
The residue is weighed as BisO*. 

B. Procedure. Isolation of Bismuth by Precipitation as the Hydroxide.^ 
The solution is taken to dryness and the residue treated with 5 cc. of nitric 
acid (1 : 4) and 25 cc. of water added. The resulting solution is poured, with 
constant stirring, into 25 cc. of concentrated ammonia and 50 cc. of 4% 
hydrogen peroxide. Upon settling of the bismuth hydroxide, the clear solution 
is filtered off and the residue is treated with more ammonia and peroxide. It 
is then filtered onto a filter paper, washed with hot, dilute ammonium hydroxide, 
(1 : 8), followed by hot water and washed free of any adhering copper or cad- 
mium (no residue when a drop is evaporated on platinum foil). Re-solution 
in hot dilute nitric acid and reprecipitation may sometimes be necessary to 
obtain the pure product. The hydroxide may be dried, ignited and weighed 
as BiaOs according to directions already given on page 66. 

Properties. BijOa'COi-HaO, mot.wt.f 526.02; sp.gr.j 6.86; decomp, by heat; 
insoluble in water, soluble in acids, insoluble in NaaCOs; white precipitate, 

Bi(0H)8, mol.wt., 259.02; loses IJ H^O at 150**; insol, in cold water and in 
alkalies; soluble in acids; white precipitate. 

Determination of Bismuth as the Sulphide, 81283 

The procedure is applicable to the determination of bismuth in a hydro- 
chloric or sulphuric acid solution, freed from other members of this group. 

Procedure. Bismuth sulphide ^ 

is precipitated by passing HjS into 
the slightly acid solution, preferably 
under pressure. When the precipi- 
tation is complete, the bismuth sul- 
phide, BisSi, is filtered off into a 
weighed Cjooch crucible, the precipi- 
tate washed with HjS water, then 
with alcohol to remove the water, 
followed by carbon disulphide to 
dissolve out the precipitated sul- 
phiur, then alcohol to remove the 
disulphide, and finally with ether. 
After drying for fifteen to twenty 
minutes, the residue is weighed as 
BiaSi. This weight multiplied by 
0.8122 =Bi. Fig. 7. — Purification of Carbon Disulphide. 

• 

Note. The carbon disulphide used should be freshly distilled. This may be 
accomplished by placing the carbon disuiphide in a small flask (A, Fig. 7) connected 
by means of a glass tube (C) to a second flask (B), cork stoppers being used. The 
vessels are immersed in beakers of water, the container with the reagent being placed 
in hot water (60-80** C.) and the empty flask in cold water. The reagent quickly dis- 
tills into the empty fla^ in pure form. 

Properties of Bismuth Sulphide. BiSiy moLwt., 512.21; sp,gr., 7-7.81; 
decomposed by heat, solubility =0.00001Sg. per 100 cc. cold HjO; soluble in nitric 
acid; brown rhombic crystals, 

* P. Jannasch, Zeit. anorg. Chem., 8, 302, 1895. 



.'Cork 





68 BISMUTH 

Determination of Bismuth as the Metal 

Reduction with Potassium Cyanide.^ Bismuth precipitated as the car- 
bonate and ignited to the oxide according to the procedure given, is fused in 
a weighed porcelain crucible with 5 times its weight of potassium cyanide over 
a low flame. The cooled melt is extracted with water, pouring the extracts 
through a filter that has been dried and weighed with the crucible. Bismuth 
is left undissolved as metallic bismuth. After washing with water, alcohol, 
and ether, the filter, with the metal and loosened pieces of porcelain glaze, is 
dried at 100° C. together with the crucible. These are then weighed and the 
increased weight taken as the amount of bismuth present in the sample. 

Electrolytic Deposition of Bismuth 

With samples containing less than 0.03 gram bismuth, the metal may be 
satisfactorily deposited by electrolysis of its sulphuric acid solution, lead having 
been removed previously by sulphuric acid by the standard procedure. The 
solution contains about 6 cc. of strong sulphuric acid per 100 cc. This is 
electrolyzed with a current of 0.6 to 0.7 ampere and about 2.7 to 3 volts 
Further details of this method may be found in " Technical Methods of Ore 
Analysis," by A. H. Low, page 54, VII Edition. John Wiley & Sons. 

VOLUMETRIC DETERMINATION OF BISMUTH 

Determination of Bismuth by Precipitation as Oxalate and 
Titration by Potassium Permanganate * 

Normal bismuth oxalate, produced by addition of oxalic acid to a nitric 
acid solution of the element, boiled with successive portions of water, is trans- 
formed to the basic oxalate. This may be titrated with potassium perman- 
ganate in presence of sulphuric acid. 

Procedure. Preparation of the Sample. One gram of the finely ground 
sample is treated with 5 to 10 cc. of concentrated nitric acid and digested on 
the steam bath and finally evaporated to dryness, the residue is taken up with 
5 cc. of nitric acid (sp.gr. 1.42)4-25 cc. of water, and diluted to 100 cc. 

Precipitation of the Oxalate. About 5 grams of ammonium oxalate or 
oxalic acid are added and the liquid boiled for about five minutes, the pre- 
cipitate allowed to settle and the supernatant solution filtered off. The 
precipitate is boiled twice with 50-cc. portions of water and the washings poured 
through the same filter. If the filtrate still passes through acid, the washing 
is continued until the acid is removed and the washing pa.*<sing through the 
filter is neutral. The bulk of the basic oxalate precipitate is placed in a beaker 
and that remaining on the filter paper is dissolved by adding 2 to 5 cc. of hydro- 
chloric acid, 1:1, the solution being added to the bulk of the precipitate. 

» Method bv H. Rose, Pogg. Ann., 110, p. 425. 

Vanino and Trcubert (Ber.. 31 (1S98), 1303), reduce bismuth by adding formal- 

tm'^o frk ita alitrVifivr n/>w1 a^Jpfir^n onrl ^Vion mnL-irtnr cfi<itiirKr ullrolirtn wifVi 1 0^o^ 




and Kyle (C. N., 76, 3). 

Muir and Robbs^ J. C. S., 41, 1. 



BISMUTH 69 

This is now wanned until it goes into solution and the liquid is diluted to 250 
cc. with hot water. Dilute ammonia is now added until the free acid is 
neutralized; the resulting precipitate is taken up with dilute sulphuric acid, 1 : 4, 
added in slight excess. The resulting solution, warmed to 70**, is titrated with 
standard potassium permanganate. 

One cc. KMnO* N/10 =0.0104 gram Bi. 

Note. Lead, copper, arsenic, iron, zinc, and tellurium do not interfere. Hy- 
drochloric acid should not be used to dissolve the sample, as it inteifeies with the 
oxalate precipitation. 

Cinchonine Potassium Iodide, Colorimetric Method ^ 

This method is applicable for the determination of small amounts of bis- 
muth, 0.00003 to 0.00015 gram, in ores and alloys. The procedure depends 
upon the fact that bismuth nitrate produces a crimson or orange color when its 
solution is added to a solution of cinchonine potassium iodide, the intensity 
of the color depending upon the amount of bismuth in the resulting product. 

Special Reagents. Cinchonine Potassium Iodide Solution. Ten grams of 
cinchonine are dissolved by treating with the least amount of nitric acid that 
is necessary to form a viscous mass and taking up with about 100 cc. of water. 
The acid is added a drop at a time, as an excess must be avoided. Twenty 
grams of potassium iodide are dissolved separately and cinchonine solution added. 
The resulting mixture is diluted with water to 1000 cc. After allowing the 
reagent to stand forty-eight hours, any precipitate formed is filtered ofif and 
the clear product is ready for use. The reagent preserved in a glass-stoppered 
bottle keeps indefinitely. It should be filtered free of suspended matter before use. 

Standard Bismuth Solution. One gram of metallic bismuth is dissolved 
in the least amount of dilute nitric acid (1:1) that is necessary to keep it in 
solution and diluted to 1000 cc, in a graduated flask. One hundred cc. of this 
solution is diluted to 1000 cc. One cc. of this diluted solution contains 0.0001 
gram bismuth. 

Procedure. Isolation of Bismuth. The solution is freed from lead by 
H1SO4, and from arsenic, antimony, and tin by precipitation of the sulphides 
and extraction with NaiS solution. The residual sulphides are dissolved in hot 
dilute nitric acid, according to the standard methods of procedure. The free 
nitric acid is nearly neutralized by the cautious addition of dilute ammonia, 
the last portion being added drop by drop, until a faint cloudiness is evident, 
and then 10 to 15 cc. of 10% ammonium carbonate aie added with constant 
stirring. The mixture is digested for about three hours on the steam bath, 
the clear solution decanted through a small filter, the residue washed by de- 
can tation once or twice with hot water containing ammonium carbonate and 
then on the filter twice with pure hot water. 

Colorimetric Comparison 

The residue of bismuth basic carbonate is dissolved in the least amount of 
dilute nitric acid necessary to effect solution and the filter washed free of bis- 
muth with a little water containing a few drops of nitric acid. The solution 
is made up to a definite volume, 50 cc. or 100 cc. according to the bulk of 

^ Method of W. C. Ferguson. 



70 BISMUTH 

precipitate dissolved. Two small beakers placed side by side may be used 
for the color comparison, a sheet of white paper or tile being placed under the 
beakers. Two 50-cc. Nessler tubes, however, are preferred. Three cc. of cin- 
chonine solution are added to each container. From a burette the bismuth 
nitrate sample is run into one of these containers in just sufficient quantity to 
color the reagent a crimson or orange tint. The exact volume required to do 
this is noted and the equivalent amount of sample used calculated. (K no 
color is produced bismuth is absent.) The reagent in the adjacent beaker or 
Nessler tube is diluted to 5 to 7 cc, and into this is run, from a burette, the 
standard bismuth nitrate solution until the color exactly matches the sample. 
From the cc. of the standard required the amount of bismuth in the sample 
can readily be calculated. 

Reaction. 3KI+C,9HaN,OKI+Bi(N03)a = CjaaNjOKIBila+SKNO,. 

Precautions. The sensitiveness of the method is lost if the depth of color is too 
great. It is neces.sary, then, to add the sample to .the cinchonine reagent in such 
quantity only as will produce a light crimson or orange color. 

Solutions in the comparison tubes or beakeis must not be overdiluted, since the 
bismuth salt forir.ed by the reaction of the cinchonine reagent is soluble in water 
with the disappearance of color in too dilute solutions. 

Comparison must be expeditiously made, as a precipitate is apt to form upon 
standing, and iodine will sometimes separate. 

The Older of addition must be observed; e.g., the bismuth solution is added to 
the cinchonine reagent, never the leverse. 

Colorimetric Determination of Bismuth. Bismuth Iodide 

Method 1 

Bismuth iodide gives an intense yellow, orange, or red color to its solu' 
tion. The color is not destroyed by SOa, as is that of free iodine. The intensity 
of the color varies as follows: 

1 part of bismuth in 10,000 parts of water produces an orange-colored solution. 
1 part of bismuth in 40,000 parts of water produces a light orange color. 
1 part of bismuth in 100,000 parts of water produces a faint yellow color. 

Reagents. Standard Bismuth Solution. One gram of bismuth is dissolved 
in 3 cc. of strong nitric acid and with 2.8 cc. of water and made up to 100 cc. 
with glycerine. Glycerine is added to keep the Bila in solution. Glycerine 
is not necessary for amounts of bismuth below 0.0075 gram per cc. 

Potassium Iodide Solution. Five grams of potassium iodide dissolved in 
5 cc. of water is diluted to 100 cc. with glycerine. 

Procedure. The sample is dissolved with just sufficient nitric acid and 
water necessary to cause solution, 10 cc. of glycerine and 10 cc. of potassium iodide 
solution added and the sample diluted to 50 cc. Comparison is now made with 
10 cc. of the standard bismuth solution to which has been added 10 cc. of 
potassium iodide and 30 cc. of water. It is advisable to have the standard 
stronger in bismuth than the sample and to draw out the standard from the 
comparison cylinder until the two colors match. 

»T. C. Thresh, Pharm. Jour., 041, 1880. 



BORON 

Wilfred W. Scott 

B f f 11 !>• l^^'^^P* ^P'O^' 2.45; ¥n.p. 2200^ ; 6.p. sub 'ime«. 
, ai.wt. 11.U, \^j^y^^ ^pg^ 2.55; m.p. 2500°; 6.p. 3500° C; oxide, B^O, 

DETECTION 

Flame Test. Boric acid is displaced from its salts by nearly all acids, 
including even carbonic acid. Upon ignition, however, it in turn drives out 
other acids which are volatile at lower temperatures. A powdered borate, 
previously calcined, is moistened with sulphuric acid and a portion placed on 
the loop of a platinum wire is heated to expel the sulphuric acid,^ then moistened 
with glycerine and placed in the colorless flame; a green color will be imparted 
to the flame. Copper salts should be removed with H2S and barium as BaSO* if 
present, as these also color the flame green. 

The flame test may be conveniently made by treating the powdered sample 
in a test-tube with sulphuric acid and alcohol (preferably methyl alcohol). A 
cork carrying a glass tube is inserted and the test-tube gently warmed. The 
escaping gas will bum with a green flame. 

The test may be made by igniting the mixture of powder, alcohol, and sul- 
phuric acid in an open porcelain dish. The green color will be seen in presence 
of a borate. The test is not as delicate as the one with the test-tube. 

Borax Bead. NajB^Or • IOH2O fused in a platinum loop, swells to several 
times its original volume as the water of crystallization is being driven out, 
then contracts to a clear molten bead. If the bead is dipped into a weak solution 
of cobalt and plunged into the flame, until it again becomes molten, the bead 
upon cooling will be colored blue. 

Turmeric Test. A few drops of acetic acid are added together with 2 
or 3 drops of an alcoholic turmeric solution to an alcoholic extract of the 
sample, placed in a porcelain dish. The solution is diluted with water and then 
evaporated to dryness on the water bath. 1/1000 milligram of boric acid will 
produce a distinct color, 2/100 milligram will give a strong reddish-brown 
colored residue, which becomes bluish-black when treated with a drop of sodium 
hydroxide solution. 

ESTIMATION 

The determination of boron is required in the valuation of borax, 
Na^A-lOHjO; boracite, 4MgB4072MgOMgCl2 ; borocalcite, CaB407-6H20; 
hydroboracite; boronatrocalcite, etc., the element being reported generally as 
the oxide, BjOi. The determination is required for obtaining the true value of 
commercial boric acid, in the analysis of fluxes and certain pigments. It is 
determined as a food-preservative in milk, meat, canned goods, etc. The ele- 
ment is determined in certain alloys of nickel, cobalt, zinc, chromium, tungsten, 
molybdenum and in the analysis of steel. 

^Silicates should be mixed with potassium fluoride and potassium acid sulphate, 
KHSO4, then held in the flame. 

71 



72 BORON 



Preparation and Solution of the Sample 

It will be recalled that crystalline boron is scarcely attacked by acids or 
alkaline solutions; the amorphous form, however, is soluble in concentrated 
nitric and sulphuric acids. Both forms fused with potassium hydroxide are 
converted to potassium metaborate. Boric acid is more readily soluble in pure 
water than in hydrochloric, nitric, sulphuric, or acetic acids, but still more soluble 
in tartaric acid (Herz, Chem. Zentr., 1903, 1, 312). It is soluble in alcohol and 
volatile oils. Borax is insoluble in alcohol. With acids it becomes transposed 
to boric acid and the sodium salt of the acid. 

Boric Oxide in Silicates, Enamel, etc. About 0.5 gram of the finely ground 
material is fused with five times its weight of sodium carbonate, the melt extracted 
with water and the extracts, containing the sodium salt of boric acid, evap- 
orated to small volume. The greater part of the excess sodium carbonate 
is neutralized with hydrochloric acid and finally made acid with acetic acid 
(litmus paper test = red). Boric oxide is now determined by the distillation 
process according to the procedure given later in the chapter. 

Boronatrocalcite, Borocalcite, Boracite, Calcium Borate. Ten grams of 
the powdered material is placed in a flask with a reflux condenser and about 50 
cc. of normal hydrochloric acid added and the mixture boiled for half an hour. 
The contents of the flask, together with the washings, including those of the 
reflux condenser (COr-free water being used), are filtered into a 500-cc. flask 
and made to volume with CO»-free water. Fifty cc. of this solution is titrated 
with half-normal sodium hydroxide, using paranitrophenol indicator. When 
a yellow color appears the hydrochloric acid has been neutralized. A second 
50-cc. portion is now taken for analysis and the free hydrochloric acid neutral- 
ized with sodium hydroxide, using the amount of caustic required in the trial 
analysis (this time without an indicator). Boric acid is now determined by titra- 
tion according to the procedure on p. 76. 

Borax, Boric Acid. Ten grams of the material are dissolved in about 
300 cc. of water (free from CO2) and made to 500 cc. in a graduated flask, with 
pure water. One hundred-cc. portions are taken for analysis and the solution 
titrated, in presence of mannitol or glycerol, according to directions given under 
the volumetric procedures. 

Boric Acid in Mineral Water. Water containing more than 0.1 gram 
boric acid per liter, about 200 cc. are evaporated to small volume, the precipitated 
salts are filtered off and washed. Boric acid passes into the filtrate and may be 
determined by the distillation method of Gooch given on p. 74. 

With water containing traces of boric acid,^ 5 liters or more are evaporated to 
about one-tenth the original volume the precipitate filtered off and washed with 
hot water. The filtrate Is evaporated down to a moist residue. If the residue is 
small, it is acidified with acetic acid and the boric acid determined by distillation, 
as stated on p. 74. If considerable residue is present, hydrochloric acid Is added 
to acid reaction, and then the mixture digested with absolute alcohol in a corked 
flask for ten to fifteen hours, with occasional shaking. The solution is filtered, the 
residue washed with 95% alcohol, the filtrate diluted with water, 10 cc. of 10% 
sodium hydroxide solution added and the alcohol distilled off. A second alcoholic 

» Reference is made to Trcadwell and Hall, Anal. Chem., 4th ed., p. 431-432 for those 
desiring more explicit details of this method. 



BORON 73 

extraction is generally recommended. The final alkaline solution is taken to 
dryness and gently ignited. The residue is extracted with water, made acid with 
acetic acid and B2O3 determined by distillation. 

Carbonates. The material is treated with sufficient acid (M. O. indicator) 
to liberate all the COi and react with the combined alkali of boric and carbonic 
acid; it is boiled in a flask with reflux condenser to expel CO2, ten to fifteen 
minutes, the solution exactly neutralized with sodium hydroxide, (M. 0.), and the 
liberated boric acid titrated in presence of glycerol and phenolphthalein as usual. 

Boric Acid in Milk, Butter, Meat and Other Foods 

Milk.^ One hundred cc. of milk is treated with 1 to 2 grams of sodium 
hydroxide, and evaporated to dryness in a platinum dish. The residue is 
thoroughly charred * by gently heating; at this stage care must be exercised or 
loss of boric acid will result; 20 cc. of water are added, the sample heated and 
hydrochloric acid added drop by drop until all but the carbon has dissolved. 
The mixture is washed into a lOO-cc. flask with as little water as possible, 0.5 
gram calcium chloride added, then a few drops of phenolphthalein indicator, 
then a 10% sodium hydroxide solution until a slight permanent pink color 
is obtained and finally 25 cc. of lime water. (All P2O8 is precipitated as calcium 
phosphate.) The liquid is made to 100 cc, mixed thoroughly, and then filtered 
through a dry filter. To 50 cc. of the filtrate, equivalent to 50 cc. of the milk 
taken, normal sulphuric acid is added until the pink color disappears, then 
methyl orange indicator is added, followed by more of the standard acid until 
the yellow color changes to a faint pink. Carbon dioxide is expelled and the 
liberated boric acid titrated in presence of glycerine, according to the procedure 
given for evaluation of borax and boric acid,. under "Volumetric Determination 
of Boron." 

Butter.' Twenty-five grams of butter are weighed out in a beaker and 25 
cc. of a sugar sulphuric acid mixture added. (Mix =6 grams sugar of milk, 
4 cc. normal sulphuric acid per 100 cc. of solution.) The beaker is placed in the 
oven (100® C.) until the fat is melted and the mixture is thoroughly stirred. 
When the aqueous solution has settled, 20 cc. are pipetted out, phenolphthalein 
added, the solution brought to boiling and half-normal sodium hydroxide added 
until a faint pink color is obtained. Ten cc. of neutral glycerine are added 
and the titration carried on until a permanent pink color appears. The dif- 
ference between the two titrations multiplied by the factor for equivalent boric 
acid gives the weight of boric acid in the portion taken. 

The determination is not affected by the phosphoric or butyric acid or by 
the sugar of milk in the butter. 

Meat.^ Ten grams of the chopped meat are mixed in a mortar with 40 to 
80 grams of anhydrous sodium sulphate, and dried in the water oven. The 
mass is powdered, then placed in a flask and 100 cc. of methyl alcohol added 
and allowed to stand for about twelve hours. The alcohol is distilled into a 
flask and saved. Fifty cc. more of alcohol are added to the residue and this 
again distilled into the first distillate. The distillates are made up to 150 cc, a 

* R. T. Thomson, Glasgow City Anal. See. Repts., 1895, p. 3. 

•The milk residue thoroughly charred will give a colorless solution upon extraction. 

• H. Droop Richmond and J. B. P. Harrison, Analyst, 27, 197. 
« C. Fresenius and G. Popp, Chem. Centr., 1897, 2, 69. 



74 



BORON 



50-cc. portion diluted with 50 cc. of water and 50 cc. of neutral glycerine added 
with phenolphthalein indicator, and the boric acid titrated with twentieth- 
normal sodium hydroxide. 

One cc. N/20 NaOH =0.0031 gram boric acid, HaBO,. 

Boric acid in canned goods, sauces, cereals, etc., may be determined by 
evaporation of the substance with sodium hydroxide and incineration as in 
case of miUc. The sodium hydroxide is neutralized and boric acid titrated as 
usual. 

QRAVIMETRIC DETERMINATION OF BORON 

The solubility of boron compounds prevents complete precipitation by any 
of the known reagents, hence most of the gravimetric methods are indirect. 

Distillation as Methyl Borate and Fixation by Lime ^ 

This excellent method, originally worked out by F. A. Gooch, ' and later modi- 
fied by Gooch and Jones,' depends upon the fact that the borates of alkaline 
earths and alkalies give up their boron in the form of the volatile methyl l)orate 
(b.p.y 65° C), when they are distilled with absolute methyl alcohol (acetone- 
free). The methyl borate passed over lime in presence of water Ls completely 

saponified, the liberated boric acid 
combining with the lime to form 
calcium borate, which mav be dried, 
ignited, and weighed. The increase 
of the weight of the lime represents 
the B2O3 in the sample. 

2B(OCH3),+CaOH-6H20 

=6CH,OHH-Ca(B02),H-3H,0. 

Procedure. About 1 gram of pure 
calcium oxide is ignited to constant 
weight over a blast lamp and then 
transferred to the dry, Erlenmeyer 
receiving flask (Fig. 8). The crucible 
in which the lime was heated and 
weighed is set aside in a desiccator for 
later use. 

0.2 gram or less of the alkali borates, 
obtained in solution by a procedure 
given under "Preparation of the Sam- 
ple," is treated with a few drops of litmus 
(or lacmoid), solution and the free al- 
kali neutralized with dilute HCl solution 
added drop by drop. A drop of dilute 
sodium hydroxide solution is added and 




\, — Distillation of Methyl Borate. 



Fig. 8. 

then a few drops of acetic acid. The slightly acid solution is transferred to the 

' Proc. Am. Acad, of ArU and Sciences, 22, 167-176 (1886). Anal. Chem., 
Treadwell-HaU, Vol. 2. 
^ bee note on p. 75. 



BORON 75 

pipette-shaped retort i?, Fig. 8, by means of the attached funnel F, washing 
out the beaker and funnel with three 2- to 3-cc. portions of water. The stop- 
cock of the funnel is closed, the apparatus is connected up as shown in the illus- 
tration, the paraffine bath, heated to not over 140** C, placed in position and 
the liquid in the retort distilled into the receiver containing the known amount 
of lime. When all the liquid has distilled over, the paraffine bath is lowered, 
the retort allowed to cool for a few minutes, 10 cc. of methyl alcohol (acetone- 
free) added to the residue in R and the contents again distilled by replacing 
the parafl^e bath. The process is repeated three times with methyl alcohol. 
The contents of the retort (which are now alkaline), are made distinctly acid 
by addition of acetic acid, and three more distillations made with 100-cc. portions 
of methyl alcohol, as before. The paraffine bath is now removed, the receiving 
flask is stoppered, the contents thoroughly mixed by shaking, and set aside for 
an hour or more for complete saponification of the methyl borate. The con- 
tents are now poured into a large platinimi dish and evaporated on the water 
bath at a temperature below the boiling-point of the alcohol. (Loss of boric 
acid will occur if the alcohol boils.) The adhering lime in the receiving flask 
is dissolved by wetting its entire surface with a few drops of dilute nitric acid 
(the flask being inclined and revolved to flow the acid over its sides). The 
contents are transferred to the dish with a little water and the evaporation 
repeated. No loss of boric acid will take place at this stage, the alcohol having 
been removed during the first evaporation. The residue is gently heated to 
destroy any calcium acetate that may have formed, the cooled borate and 
lime are taken up with a little water and transferred to the crucible in which 
the lime was heated and weighed. The material clinging to the dish is dis- 
solved with a little nitric acid (or acetic acid), and washed into the crucible. 
The contents of the crucible are evaporated to dryness on the water bath, then 
heated very gently over a flame (the crucible being covered) and finally more 
strongly. The heating is continued until a constant weight is obtained. The 
increase of weight of the lime represents the amount of BjOj in the sample. 

Notes. Gooch and Jones worked out a procedure which utilizes sodium tun^- 
state as a retainer of the methyl borate, in place of the lime. This substance is 
definite in weight, not hydroscopic, soluble in water, and recoverable in its original 
weight after evaporation and ignition. ^'Methods in Chem. Anal.," p. 204, 1st Ed. 
By F. A. Gooch, John Wiley & Sons, Publishers. 

The receiving flask has a cork stopper with a hole to accommodate the tube of 
the condenser and a slit to permit the escape of air from the flask. 

Gooch reconmiends coolmg of the receiving flask. 



76 BORON 



VOLUMETRIC DETERMINATION OF BORON 

Titration of Boric Acid in Presence of Mannitol or Glycerol 

Evaluation of Borax 

The method takes advantage of the fact that boric acid reacts neutral to 
methyl orange (or paranitrophenol), but is acid to phenolphthalein, and may 
be quantitatively titrated in the presence of mannitol or of glycerol, which 
prevent the hydrolization of sodium borate. If insufficient mannitol or glycerol 
are present the color change takes place too soon, the color fading upon adding 
more of these substances. The end-p>oint is reached when the further addition 
of these reagents produces no fading of the color. In the procedure, the alkali 
is neutralized in presence of methyl orange (or paranitrophenol), and the liberated 
boric acid is now titrated. 

Reactions. Na,B407H-2HCl-|-5H,0 =2NaCl-h4H3BO, 
HsBOs-hNaOH =NaB0,+2H,0. 

Procedure. One hundred cc. of the solution containing the borax, pre- 
pared according to directions under "Preparation and Solution of the Sample," 
equivalent to 2 grams of the substance, is taken for analysis. 

A. Titration of Combined or Free Alkali. Methyl orange indicator is 
added and the solution is titrated with normal or half-normal sulphuric acid 
until the yellow color is replaced by an orange red. (With paranitrophenol the 
solution becomes colorless.) From this titration the combined alkali, together 
with any free alkali, is calculated. If free alkali is known to be absent (see 
note), the amount of borax may be calculated. 

One cc. N. HjSO* =0.031 gram Na,0, or =0.1911 gram Na2B407 • 10H,O, 
or =0.101 gram NajBiOi. 

B. Titration of Boric Acid. The liberated boric acid may now be titrated 
with caustic. This may be accomplished either on the above portion or on a 
fresh 100-cc. portion (free from methyl orange indicator), to which the amount 
of acid, required to neutralize the alkali, has been added. Fifty cc. of neutral 
glycerol or 1 gram of mannitol are added, followed by phenolphthalein indicator. 
Normal or half-normal sodium hydroxide is added from a burette until a change 
of color takes place. If methyl orange is present, the color, first becoming yellow, 
changes to an orange red. In absence of methyl orange the characteristic 
lavender or purplish pink of alkali phenolphthalein is obtained. More glycerol 
or mannitol is now added and if the color fades the titration is continued until 
the addition of these reagents no longer produces this fading of the end-point, 
From this titration boric acid is calculated and the equivalent borax determined. 

One cc. N. NaOH =0.062 gram HaBOj, equivalent to 0.0505 NajBA, or 
0.0955 NajBiOylOHiO. 

Factors. Na^O to Na^BA =3.2581 , reciprocal =0.3069. 

NaaO to NajBA-lOHsO =6.1638, recip. =0.1622. 

Na,0 to NajCO, = 1.7007, recip. =0.5849. 

Na,0 to Na^COalOHoO =4.6155, recip. =0.2167. 

Notes. In borax (free from excess BjOj or NaaO), the acid titration is half the 
subsequent alkali titration (factor, acid to borax =0.1911, alkali to borax =0.0955). 



BORON 77 

If the acid titration exceeds this proportion, alkali other than that combined with 
boric acid is indicated; if the alkali titration is greater than twice the acid titration, 
free boric acid is indicated. 

The glycerol should be made neutral with N/10 NaOH before use, in case it con- 
tains free fatty acids. 

Mannitol is a solid and has some advantages over glycerol; it gives a shari>er 
end-point, is less apt to contain free acids and does not appreciably alter the bulk 
of the solution to which it is added. (L. C. Jones, C. N., 80, 65, 1899). 

N.B. Paper on use of mannitol and glycerol in determining boric acid, by II. 
T. Thomson, J. S. C. I.. 12, 432. 

Examjde. By actual test 2 grams NaiB^OT'lOHjO required 10.66 cc. N. H2S04 = 
10.66X0.1911=2.037 gram NajBiOi-lOHjO. The liberated HaBO, required 21.39 
cc. of N. NaOH=21.39X0.0955 = 2.043 gram Na2B407lOHtO. The borax had lost 
a small amount of water of crystallization, hence the high results when calculated to 
Na,B4O7l0HiO. 

EVALUATION OF BORIC ACID 

One hundred cc. of the solution, prepared as directed under "Preparation 
of the Sample," equivalent to 2 grams of the original material, is treated with 
50 cc. of glycerol or 1 gram of mannitol, and the acid titrated with standard 
caustic, in presence of phenolphthalein indicator according to the procedure 
given in By under "Evaluation of Borax." 

One cc. normal acid contains 0.062 gram H»BOj, hence the cc. of caustic 
required multiplied by 0.062 = grams boric acid. 

Examples, Two grams H»BOa by actual test required 32.1 cc. N. NaOH 
=32.1 X .062 = 1.99 grams H,BOs. 

Detection of Minute Amounts of Boron. 

Robin's Test for Boron. To a few drops of the aqueous solution under examination 
(slightly acidified with HCl) are added two drops of a tincture of mimosa flowers, and 
the mixture evaporated to dryness on the water bath. The residue is treated with 
dilute ammonia water j whereupon in presence of boric acid, a rose pink to blood red 
color develops, accordmg to the amount present. L. Robin claims that as little as 
0.0001 milligram may be detected in presence of nitrates, chlorides, iodides, or calcium 
sulphate. Organic acids and sodium phosphate interfere. The reagent is prepared by 
extracting the mimosa flowers with ethyl alcohol. The extract is protected from the 
light. 



BROMINE 

Wilfred W. Scott 

Br, at.wt. 79.92; 9p.gr. 3.1883''; m.p. —7.3''; b.p. 58.7"" C; acids, HBr, 

HBrO, HBrOs 

DETECTION 

Silver Nitrate solution precipitates silver bromide, AgBr, light yellow, 
from solutions containing the bromine anion. The precipitate is insoluble 
in dilute nitric acid, but dissolves with difficulty in ammonium hydroxide 
and is practically insoluble in ammonium carbonate solution (distinction from 
AgCl). 

Carbon Disulphide or Carbon Tetrachloride shaken with free bromine 
solution, or with a bromide to which a little chlorine water has been added, 
(a large excess of chlorine must be avoided, as this forms BrCl compound), will 
absorb the bromine and become a reddish-yellow color, or if much bromine 
is present, a brown to brownish-black. In the latter case a smaller sample 
should be taken to distinguish it from iodine. 

Bromates are first reduced by a suitable reducing agent such as cold oxalic 
acid, sodium nitrite, hydrochloric acid, etc., and the liberated bromine tested 
as directed above. Silver nitrate added to bromates in solution precipitates 
AgBrOi, which is decomposed by hydrochloric acid to bromine gas. 

Barium Chloride precipitates Ba(BrO»)i, which is reduced readily to bromine 
as directed above. 

Magenta Test for Bromine.^ The test reagent is made by adding 10 cc. 
of 0.1% solution of magenta to 100 cc. of 5% solution of sulphurous acid and 
allowing to stand until colorless. This is the stock solution. Twenty-five cc. 
of this reagent is mixed with 25 cc. of glacial acetic acid and 1 cc. of sulphuric 
acid. Five cc. of this is used in the test. 

Test. Five cc. of the magenta reagent is mixed with 1 cc. of the solution 
tested. Chlorine produces a yellow color. Bromine gives a reddish-violet 
coloration. The colored compound in each case may be taken up with chloro- 
form or carbon tetrachloride and a colorimetric comparison made with a 
standard. 

In halogen mixes, iodine is first eliminated by heating with an iron per 
salt. Broipine is now liberated by adding sulphuric acid and potassium chromate. 
A glass rod with a pendant drop of sodium hydroxide is held in the vapor to 
absorb bromine, and the drop then tested with the magenta reagent. After 
iodine and bromine are eliminated, chlorine may be tested by heating the sub- 
stance with potassium permanganate, which liberates this halogen. 

^G. Denig^ and L. Chelle. Ann. Chira. anal., 1913, 18, 11-15; The Analyst, 
1913, 119. 

78 



BROMINE 79 



ESTIMATION 

Bromine never occurs free in nature. It is found chiefly combined with 
the alkalies and the alkaline earths, hence occurs in many saline springs and 
is a by-product of the salt industry. It is found in silician zinc ores, Chili 
saltpeter, in sea water (probably as MgBrs), in marine plants. Traces occur in 
coal, hence in gas liquors. 

The substance is used in metallurgy, the arts, and medicine. It is a valu- 
able oxidizing agent for the laboratory. 

Preparation and Solution of the Sample 

The following facts regarding solubility should be remembered: The ele- 
ment bromine is very soluble in alcohol, ether, chloroform, carbon disulphide, 
carbon tetrachloride, concentrated hydrochloric acid and in potassium bromido 
solution. One hundred cc. of water at 0° C. is saturated with 4.17 grains of 
bromine, and at 50° C. with 3.49 grams. The presence of a number of salts 
increases its solubility in water, e.g., BaCU, SrClj, etc. 

Note. The element is a dark, brownish-red, volatile liquid, giving ofT a dark reddish 
vapor with suffocating odor, irritating the mucous membrane (antidote dil. NH4OH, 
ether), very corrosive. Acts violently on hydrogen, sulphur, phosphorus, arsenic, 
antimony, tin. the heavy metals, and on potassium, but has no action on sodium, 
even at 200° C. Bleaches indigo, litmus, and most organic coloring matter. It is a 
strong oxidizing agent. Bromme displaces iodine from its salts, but is displaced by 
chlorine from its combinations. 

Bromides are soluble in water, with the exception of silver, mercury, lead, 
and cuprous bromides. 

Bromates are soluble in water with the exception of barium and silver bro- 
mates and some basic bromates. 

Decomposition of Organic Matter for Determination of Bromine. The 
substance is decomposed with nitric acid in presence of silver nitrate in a bomb 
combustion tube by the Carius method described in the chapter on Chlorine, 
under "Preparation and Solution of the Sample " The residue, containing 
the halides, is dissolved in warm ammonia water, and filtered, as stated. The 
filtrate and washings are acidified with nitric acid, heated to boiling and the 
silver bromide settled in the dark, then filtered through a weighed Gooch cru- 
cible, the washed precipitate dried at 130° C. and weighed as AgBr. 

In presence of two or three halogens the lime method is recommended, as 
given in the chapter on chlorine, page 122. 

Salts of Bromine. The ready solubility of bromides and bromates has been 
mentioned. A water extract is generally sufficient. Insoluble salts are decom- 
posed by acidifying with dilute sulphuric acid and adding metallic zinc. The 
filtrate contains the halogens. 

SEPARATIONS 

Separation of Bromine from the Heavy Metals. Bromides of the heavy 
metals are transposed by boiling with sodium carbonate, the metals being pre- 
cipitated as carbonates and sodium bromide remaining in solution. 

Separation of Bromine from Silver (AgBr) and from Cyanides (AgCN). 
The silver salts are heated to fusion. The mass is now treated with an excess 



80 BROMINE 

of zinc and sulphuric acid, the metallic silver and the paracyanogen filtered 
ofT and the bromine determined in the filtrate. 

Separation of Bromine from Chlorine or from Iodine. Details of the 
procedure for determining the halogens in presence of one another is given in 
the chapter on Chlorine, page 130. Free bromine is liberated when the solu- 
tion of its salt is treated with chlorine. 

Separation of Bromine from Iodine.^ The neutral solution containing 
the bromide and iodide is diluted to about 700 cc. and 2 to 3 cc. of dilute sul- 
phuric acid, 1:1, added, together with about 10 cc. of 10% sodium nitrite, 
NaNOj, solution. (Nitrous acid gas may be passed through the solution in 
place of adding sodium nitrite, if desired.) * The solution containing the halides 
is boiled imtil colorless and about twenty minutes longer, keeping the volume 
of solution above 600 cc. 0.5 gram KI may be decomposed and the iodine 
expelled from the bromide in half an hour. The bromine is precipitated from the 
residue remainiag in the flask by addition of an excess of silver nitrate and 
determined as silver bromide. 

The procedure for determining iodine is given in the chapter on this subject. 



GRAVIMETRIC METHODS 
Precipitation as Silver Bromide 

The general directions for determination of hydrochloric acid and chlorides 
apply for determining hydrobromic acid and bromides. 

I. Hydrobromic Acid and Bromides of the Alkalies and Alkaline Ear^s. 

Procedure. The bromide in cold solution is made slightly acid with nitric 
acid and then silver nitrate added slowly with constant stirring until a slight 
excess is present. The mixture is now heated to boiling and the precipitate 
settled in the dark, then filtered through a weighed Gooch crucible, and washed 
with water containing a little nitric acid and finally with pure water to remove 
the nitric acid. After ignition the silver bromide is cooled and weighed as AgBr. 

AgBrX 0.4256= Br, or X0.6337=KBr. 

n. Heavy Metals Present. 

If heavy metals are present it is not always possible to precipitate silver 
bromide directly. The heavy metals may be removed by precipitation with 
ammonia, sodium hydroxide or carbonate and the bromide then determined 
in the filtrate as usual. 

VOLUMETRIC METHODS 

Free hydrobromic acid may be titrated with standard alkali exactly as is 
described for the determination of hydrochloric acid in the chapter on Acids. 
One cc. normal caustic solution is equivalent to 0.08093 gram HBr. 

^F. A. Gooch and J. R. Ensign. Am. Jour. Sci., (3), xl, 145. 

* Nitrous acid gas is generated by dropping dilute H2SO4, by means of a separatory 
funnel onto sodium nitrite in a flask. 



BROMINE 81 



Determination of Free Bromine. Potassium Iodide Method 

The method depends upon the reaction KI+Br=KBr+I. 

Procedure. A measured amount of the sample is added to an excess of 
potassium iodide, in a glass-stoppered bottle, holding the point of the delivering 
burette just above the potassium iodide solution. The stoppered bottle is 
then well shaken, and the liberated iodine titrated with standard thiosulphate 
solution. 

One cc. of N/10 thiosulphate, Na^StO, =0.007992 gram Br. 

Determination of Bromine in Soluble Bromides. Liberation of 

Bromine by Addition of Free Chlorine 

When chlorine is added to a colorless solution of a soluble bromide, bromine 
is Uberated, coloring the solution yellow. At boiling temperature the bromine is 
volatiUzed, the liquid becoming again colorless. When the bromide is completely- 
decomposed and bromine expelled, further addition of chlorine produces no color 
reaction. KBr+Cl=KCl+Br. 

Procedure. The solution containing the bromide is heated to boiling and 
standard chlorine water added from a burette (protected from the light by 
being covered with black paper), the tip of the burette being held just above 
the surface of the hot bromide solution to prevent loss of chlorine. The reagent 
is added in small portions until finally no yellow coloration is produced. From 
the value per cc. of the chlorine reagent the bromine content is readily calculated. 

Standard Chlorine Water. The reagent is made by diluting 100 cc. of 
water saturated with chlorine to 500 cc. This solution is standardized against 
a known amount of pure potassium bromide (dried at 170® C), the same 
amount of bromide being taken as is supposed to be present in the solution 
examined. The value per cc. of the reagent is thus established. 

Silver-Thiocyanate-Ferric Alum Method. (Volhard) 

The procedure is the same as that used for the determination of chlorine. 
The bromide solution is treated with an excess of tenth-normal silver nitrate 
solution, and the excess of this reagent determined by titration with ammonium 
thiocyanate, using ferric alum indicator. One cc. of the thiocyanate should be 
equivalent to 1 cc. of silver nitrate solution. The formation of the red ferric 
thiocyanate indicates the completed reaction. (Consult the procedure in the 
chapter on Chlorine, page 125.) 

One cc. of N/10 AgNO, =0.007992 gram Br. 

Determination of Traces of Bromine 

By means of the magenta reagent, described under "Detection," small 
amoimts of bromine may be determined colorimetrically. 

To 5 cc. of the solution is added 0.2 cc. of strong hydrochloric acid, 1 cc. 
of concentrated sulphuric acid, 1 cc. of the stock magenta reagent and 0.2 cc, 
of a 10% solution of potassium chromate, shaking the mixture with additioF 



82 



BROMINE 



of each reagent, and without cooling, 1 cc. of chloroform is added. Comparison 
is made with a standard sample containing a known amount of bromide.^ 

Note. A solution containing 0.001 gram bromine per liter has a violet to reddish- 
violet color. 



Determination of Bromates by Reduction with Arsenous Acid 

and Titration of tlie Excess > 

Bromic acid may be reduced by arsenous acid in accordance with the reac- 
tion 3H»As08-|-HBrOj=3H^As04+HBr. In the process a considerable excess 
of ai*senous acid is added, the excess titrated with iodine and the bromate 
calculated. 

Procedure. The sample of bromate, dissolved in water, is treated with a 
considerable excess of N/10 arsenous oxide (dissolved in alkali hydrogen car- 
bonate) reagent, the solution then acidified with 3 cc. to 7 cc. of dilute sul- 
phuric acid (1 : 1) and diluted to a volume not exceeding 200 cc. After boiling 
for ten minutes, the free acid is neutralized with alkali hydrogen carbonate 
(NaHCOa or KHCO,) and the excess of arsenite titrated with N/10 iodine. 

Let X cc. equal the difference between the two titrations with N/10 iodine (i.e. 
of total arsenite minus excess arsenite) and w equal the weight of bromate de- 
sired, then 



w 



-C 



X cc.Xmol. wt. RBrO 



6X10X1000 



-jmilli 



igrams. 



ANALYSIS OF CRUDE POTASSIUM BROMIDE AND 

COMMERCIAL BROMINE 

Determination of Chlorine, Combined or Free 

This is the principal impurity present and its estimation is concerned here. 
Andrews' modification of Dugarszk's method • is as follows: 

Procedure. The following amount of sample and reagents should be 
taken. 



Approx. ppr cent Impurity 
if KCl Present is 


Amount Substance to 
be Taken, Granu 


lodate Solution 1/5 N. 
Required: cc. 


iN. HNOi Required, 
cc. 


Over 5 
1.5 to 5 
0.2to 1.5 


0.6 
1.8 
3.6 


36 

96 

186 


20 
26 
35 



» O. Dcnig<58 and L. Chelle, Ann. Chem. anal., 1913, 18-15; Analyst, 1913, p. 119. 
By means of the magenta reagent it is possible to detect bromine in the ash of 

Slants, beet root, spinach, etc. The organic substance may be decomposed by 
eating in a combustion tube. Filter paper moistened with the reagent and held in 
the fumes of the organic substances gives the characteristic test if bromine is present. 
« Method of F. A. Gooch and J. C. Blake, Am. Jour. ScL, 14, Oct., 1902. Pro- 
cedure communicated to the Editor by Prof. Gooch. 

» Jour. Am. Chem. Soc., 1907, 29, 275-283; Zeits. anorg. Chem., 1895, 10, 387. 



BROMINE 83 

The mixture is gently heated to boiling in a long-necked Kjeldahl flask, 
inclined at an angle of 30°, potassium iodate solution added, then nitric acid 
and sufficient water to make the volume about 250 cc. The boiling is con- 
tinued until bromine is expelled (test steam with 2% KI solution rendered 
faintly acid with hydrochloric acid). The mixture is boiled down to not below 
90 cc. Now 1 to 1.5 cc. of 25% phosphorus acid are added and the mixture 
boiled for five minutes after all the iodine has been expelled. The colorless 
Uquid is cooled, mixed with a slight excess of 1/20 or 1/50 normal silver nitrate 
solution (according to the proportion of chloride), the excess of silver nitrate 
then determined by titration with standard thiocyanate with ferric nitrate as 
indicator. (See procedure for silver-thiocyanate-ferric alum method of Volhard 
for determination of chlorine, page 125.) 

Determination of Chlorine in Crude Bromine 

Three grams of bromine (or more if less than 0.5% chlorine is present) in 
50 cc. of 4% potassium iodide solution in a glass-stoppered flask (cooled in ice 
during hot weather) are shaken and then transferred to a Kjeldahl flask. Sixty 
cc. of 1/5 N. KIOj solution and 24 cc. 2N. HNOj introduced, the solution 
diluted to 250 cc. and chlorine determined as directed above. 



CADMIUM 

Wilfred W. Scott 
Cd^ at.wt. 112.4; ap.gr. 8.642; m.p. 320.9'' M b.p. US'" C; oxide, CdO 

DETECTION 

Cadmium is precipitated by hydrogen sulphide from an acid solution as 
yellow cadmium sulphide, CdS. The precipitate is insoluble in ammonium 
sulphide (distinction from arsenic, antimony, and tin), but dissolves upon 
addition of hot nitric acid (separation from mercury). Upon addition of sul- 
phuric acid and expulsion of nitric by taking the solution to SOj fumes, and 
dilution with water, cadmium remains in solution (lead is precipitated, PbS04). 
Bismuth is precipitated by ammonium hydroxide and removed by filtration. 
Potassium cyanide is added to prevent the precipitation of copper sulphide; 
and hydrogen sulphide is led into the solution, whereupon cadmium precipitates 
as yellow CdS. 

Cadmium gives a brilliant spectrum of green and blue lines. 

Blowpipe Tests. Heated on charcoal in the reducing flame, cadmium gives 
a brown incrustation. The residue is volatile in the reducing flame. 

ESTIMATION 

The element occurs combined as the sulphide in small quantities. In the 
mineral greenockite it occurs as the principal element. As it occurs in prac- 
tically all zinc ores and is found in most coimnercial zinc, it is determined in 
the analysis of these substances. It is a by-product of lead and zinc smelting. 
The element is determined in certain alloys, especially those used for trial plates 
for silver coinage. It is determined in paint pigments; e.g., CdS, yellow. 

Preparation and Solution of the Sample 

The metal is slowly soluble in hot, moderately dilute hydrochloric acid or 
sulphuric acid, nmch more readily in nitric acid. It is soluble in ammonium 
nitrate. The oxide is readily soluble in acids. 

Treatment of Ores 

Sulphides are best dissolved by treating 0.5 to 1 gram of the finely powdered 
ore with 15 to 20 cc. of strong hydrochloric acid and 10 cc. of strong nitric 
acid. After standing on the water bath for ten to fifteen minutes, the solution 
is boiled until the sulphides are decomposed, additional hydrochloric being 
added if necessary. Unless silica is known to be al)sent the solution is taken 
to dryness and the residue dehydrated in the air oven for an hour. Five to ten 
cc. of strong hydrochloric acid and about 25 cc. of water are added and the 

iCir. 35 (2d Ed.), U. S. Bureau of Standards. 

84 



CADMIUM 85 

mixture heated to boiling. The residue of silica should appear white. This 
is filtered off and cadmium determined in the filtrate after making the necessary 
separations. 

If lead is present it is advisable to add 5-6 cc. of concentrated sulphuric 
acid to the cooled solution after the hydrochloric-nitric acid treatment and 
to evaporate to SOj fumes. After cooling, 50 cc. of water are added and the 
mixture heated to boiling, then placed on the steam bath until any iron present 
has completely dissolved. Silica and lead are now filtered off and the filtrate 
treated as directed under *' Separations." 

Carbonates may be dissolved by hydrochloric acid alone. Evaporation 
to dryness is necessary if silica is present. 

Alloys are best dissolved in hydrochloric and nitric acids, followed by addition 
of sulphuric acid, and nitric acid then expelled by evaporating the solution 
to SOi fumes. 

SEPARATIONS 

Removal of Silica. The procedure has been given under "Preparation 
arid Solution of the Sample." 

Separation from the Ammonium Sulphide Group, the Alkaline Earths 
and the Alkalies. The solution, acidified with 2 cc. of concentrated sulphuric 
acid or about 5 cc. of strong hydrochloric acid per 100 cc, is treated with 
hydrogen sulphide to saturation. The precipitate, containing cadmium sul- 
phide with other members of the group that were present in the original solu- 
tion, is filtered off and washed with hydrogen sulphide water slightly acidulated 
with hydrochloric acid. 

Removal of Arsenic, Antimony, and Tin. Treatment in Absence of 
Copper. The precipitate is rinsed from the filter into the beaker as completely 
as possible with no more water than is necessary. The beaker is placed under 
the filter and cold solution of potassium hydroxide (20%) is poured onto the 
filter. (Sodium hydroxide will do.) Arsenic, antimony, and tin will dissolve 
and leave cadmium sulphide. A dark-colored residue indicates the presence 
of bismuth, lead, and less frequently, of mercury. 

Treatment in Presence of Copper. A strong solution of potassium cyanide 
may be used in place of a fixed alkali hydroxide. By this treatment, the copper 
is removed along with arsenic, antimony, and tin. 

If the precipitate is yellow or orange-colored,^ it is dissolved in hydrochloric 
acid, after thorough washing with hydrogen sulphide water, and the solution 
treated according to one of the procedures given later. 

Removal of Lead and Bismuth. Should the above precipitate appear 
dark-colored, lead, bismuth, and possibly mercury are indicated. In the pre- 
liminary treatment of the ore with sulphuric acid, the lead is generally com- 
pletely removed as lead sulphate, but traces may be present in the filtrate. The 
moist precipitate and filter are placed in a flask and 10 cc. of strong hydro- 
chloric acid added, with an equal amount of water. The mixture is boiled until 
the cadmium sulphide dissolves, the HjS gas being driven out of the solution. 
The solution, diluted with 25 cc. of water, is filtered, and the filter washed with 
hot water. Any dark residue may be rejected. The filtrate is diluted some- 
what and then sodium carbonate added in slight excess, followed by 1 oi^ 2 
grams of potassium cyanide. After digesting for some time at a gentle heat 

^ Cadmium sulphide precipitated from a sulphuric acid solution is orange-colored. 



86 CADMIUM 

the mixture is filtered and washed with cold water. Bismuth and lead remain 
on the filter as carbonates. HsS is now passed into the filtrate, diluted if 
necessary. This should precipitate pure cadmium sulphide, unless mercury 
is present. The residue is washed with hydrogen sulphide water, and then 
dissolved in hydrochloric acid. 

Separation of Cadmium from Mercury. This separation is seldom re- 
quired. The procedure is based upon the insolubility of mercury sulphide in hot 
dilute nitric acid, whereas cadmium sulphide is readily soluble. The two sul- 
phides are boiled with nitric acid, 1 : 3, filtered and the residue washed with hot 
water. The filtrate is evaporated with a little sulphuric acid to small volume 
on the hot plate and then to SOs fumes. (Spattering during the last stages of 
removal of water will cause loss unless the recepticle is covered.) The cooled 
residue is taken up with water and if any insoluble matter remains it is filtered 
off. Cadmium is now determined in the solution. 



GRAVIMETRIC METHODS FOR THE DETERMINATION OF 

CADMIUM 

Determination as Cadmium Sulphate, CdS04 

The hydrochloric acid solution of cadmium obtained under the section on 
isolation of the cadmium is evaporated to dryness on the water bath in a weighed 
platinum crucible or dish. The residue is covered with a slight excess of dilute 
sulphuric acid, the solution again evaporated as far as possible on the water 
bath, and finally the excess sulphuric acid expelled by gently heating. This 
final stage is best accomplished by placing the crucible in a larger one, pro- 
vided with an asbestos ring to separate the two. The outer crucible may now 
be heated to redness without danger of decomposing the cadmium sulphate. 
The heatang is continued until no more fumes of sulphuric acid are evolved. 
The residue is weighed as cadmium sulphate, CdS04. 

CdSO4X0.5392=Cd. 

Electrolytic Determination of Cadmium 

This method for determination of cadmium is exceedingly accurate. The 
procedure recommended by Treadwell * gives excellent results. 

Procedure. A drop of phenolphthalein is added to the cadmium sulphate 
solution (obtained by evaporating the hydrochloric acid solution with sulphuric 
acid to SOs fumes), then a solution of pure caustic soda until a permanent red 
color is obtained. A strong solution of potassium cyanide is now stirred in, adding 
drop by drop, until the cadmium hydroxide precipitate just dissolves (an excess 
should Ixj avoided). The solution is diluted to alx)ut 100 cc. with water and 
electrolyzed in the cold, usitig a gauze cathode, the current being 0.5 to 0.7 ampere 
and the electromotive force 4.8 to 5 volts. At the end of five or six hours the 
ciyrent is iticreased to 1-1.2 ainixjrcs, and the solution electrolyzed for an hour 
more. 

^ Treadwell and Hall, Analytical Chem., Vol. II Beilstein and Jawein, Ber., 12, 
446. 



CADMIUM 87 

The liquid is quickly poured off, or better, the beaker lowered, and another 
of water substituted. The deposited metal is then washed by dipping the 
cathode in alcohol and finally in ether. After drying at 100° C., the cooled 
cathode is weighed. The increase of weight represents the deposited metal, 
cadmium. 

E. F. Smith, ^ reconunends the addition of one gram of potasj^ium C3anicle to 
50 cc. solution of the chloride or sulphate salt, followed by dilution to 125 cc. The 
electrolysis is conducted at a temperature of 60** C. with N.D.ioo = .06 ampere and 
E.M.F. =3.2 volts. 

Rapid deposition can be effected by means of the rotating anode (600 revolutions 

per minute). The solution of cadmium sulphate containing 3 cc. of HSOi (1 : 10) 

per 150 cc. The solution, heated to boiUng, is electrolyzed with a current of N.D.;oo 

= 5 amperes, E.M.F. =8-9 volts. Fifteen minutes is sufficient for the deposition 

of .5 gram of cadmium. 

Notes. Before washing and discontinuing the current, it is advisable to add 
a little water to raise the level of the li<)uid and continue the electrolysis to ascertain 
whether the deposition is complete. 

Traces of cadmium may be estimated in the above solution by saturating this 
with HtS and comparing the yellow-colored colloidal cadmium sulphide solution with 
a known quantity of cadmiiun and the same amounts of potassium hydroxide and 
cyanide as in the solution tested. 

VOLUMETRIC DETERMINATION OF CADMIUM 
Titration of Cadmium Sulphide with Iodine.^ 

The titration of cadmium sulphide with standard iodine in a hydrochloric 
acid solution is the same as the procedure given for determination of sulphur 
by the evolution method, the following reaction taking place: 

CdS+2HCl+l2 =CdCla+2HI+S. 

Procedure. Cadmium having been isolated as the sulphide according to 
the standard procedures given, the precipitate is washed and allowed to drain 
on the filter. The filter, together with the sulphide, is placed in a beaker 
or an Erlenmeyer flask, water added, and the whole shaken to break up the 
precipitate. A moderate quantity of hydrochloric acid is added and the solu- 
tion titrated with standard N/5 or N/10 iodine solution. Towards the end 
a little starch solution is added and the titration continued until the excess 
of iodine colors the solution blue. If preferred, an excess of iodine solution 
may be added and the excess determined by a back-titration with standard 
thiosulphate solution. 

One cc. N/10 iodine =0.00532 gram cadmium. 

> Electro-Analysis, E. F. Smith. P. Blakiston's Son & Co. Pub. 

* P. yon Berg (Z. a. C, 26, 23) transfers the precipitate and filter to a stoppered 
flask, expels the air with COj and by boiling and then titrates in an hydrochloric acid 
solution. Experiments by the author have shown this caution to be unnecessary. 



CALCIUM 

Wilfred W. Scott 
CsL^at.wt. 40.07; 8p.gr. 1.5446^®*; m.p. 810° ^C; oxide, CaO 

DETECTION 

In the usual course of qualitative and quantitative analysis calcium passes 
into the filtrates from the elements precipitated by hydrogen sulphide in acid 
and alkaline solutions (Ag, Hg', Hg", Pb, Cu, Cd, As, Sb, Sn, Fe, Cr, Al, Mn, 
Ni, Co, Zn, etc.), and is precipitated from an ammoniacal solution by am- 
monium carbonate as calcium carbonate, along with the carbonates of barium 
and strontium. The separation of calcium from barium and strontium is con- 
sidered under Separations. The oxalate of calcium is the least soluble of 
the alkaline-earth group.' All, however, are soluble in mineral acids. Calcium 
oxalate may be precipitated from weak acetic acid solution by ammonium 
oxalate. 

Flame Test. The flame of a Bunsen burner is colored vellowish red when 
a platinum wire containing calcium salt moistened with c ncentrated hydrochlo- 
ric acid is held in the flame. 

Spectrum. An intense orange and green line with a less distinct violet 
line. Note chart of the spectra of the alkaline earths. Plate II. 

See chapter on Barium under Separations — Preliminary Tests, page 52. 

ESTIMATION 

The determination of calcium is required in complete analyses of ores. It 
is of special importance in the analysis of mortar, cement, bleaching powder, 
plaster of Paris, certain paint pigments such as phosphorescent paint, CaS. 
The determination is required in the analysis of water. 

Calcium occurs in the following substances: as carbonate in limestone, 
marble, chalk, Iceland spar, shells, coral, pearl. Together with magnesium it 
is found in dolomite. It occurs as sulphate in anhydrite, gypsum, alabaster, 
selenite; as silicate in the mineral wollastonite, CaSiOs; as phosphate in phos- 
phorite, Ca3(P04)2, also in bones and in apatite, 3Ca3(P04)2CaF2; as fluoride 
in fluorspar, CaF2. As oxalate it occurs in plant cells. It is found in nearly all 
mineral springs, artesian wells, and river waters, principally as bicarbonate of 
calcium, CaHCOj. 

Preparation and Solution of the Sample 

The oxide, hydroxide, and salts of calcium are soluble in acids with the 
exception of gypsum and certain silicates which require fusion with sodium 
carbonate or bicarbonate followed by an hydrochloric acid extraction. 

»Cir. 35 (2d Ed.) U. S. Bureau of Standards. 

"Solubility: CaC204H,0 = 0.000554 gram per 100 cc. H.O. BaCjOi-HjO =0.0093 
gram. SrCO* 11,0=0.0051 gram. MgC,042H20 = 0.07 gram. 
Van Nostrand's Chem. Annual — Olsen. 

88 



' CALCIUM 89 

Solution of Limestones, Dolomites, Magnesites, Cements, Lime, etc. 
One gram of the powdered material is digested in a 250-cc. beaker with 
20 cc. of water, 5 cc. of concentrated hydrochloric acid, and 2 or 3 drops of 
nitric acid (sp.gr. 1.42). The beaker is covered to prevent loss by effervescence. 
When the violent action has subsided, the sample is placed on a hot plate and 
boiled for a few minutes. The watch-glass is rinsed into the beaker and the 
solution filtered. The residue is washed, dried and ignited in a platinum cru- 
cible, and then fused with a little sodium carbonate or bicarbonate. The cooled 
fusion is dissolved in hot dilute hydrochloric acid, the liquid added to the main 
solution and calcium determined by precipitation as calcium oxalate, after removal 
of silica, iron, alumina, etc. 

Solution of Gypsum, Plaster of Paris, and Sulphates of Lime, etc. The 
treatment of the sample is similar to the one given above with the exception 
that it is advisable to add a larger amount of strong hydrochloric acid, e.g., 
about 20 to 25 cc. If barium sulphate is present it is indicated by the clouding 
of the solution, upon acidifying the water extract of the carbonate fusion. 

Silicates. Solution of silicates is best obtained by direct fusion of 1 gram 
of the powdered material with 4 to 5 grams of sodium carbonate, in a plati- 
num crucible. The cooled melt is now covered with water and dissolved with 
hydrochloric acid according to the standard procedure for carbonate fusions. 
The hydrochloric acid solutions are taken to dryness and the silica dehydrated 
in an oven at 110° C. for an hour and then the residue is extracted with dilute 
hydrochloric acid and filtered. The filtrate contains iron, alumina, magnesium, 
lime, etc. 

Chlorides, Nitrates, and Other Water-soluble Salts. These are dissolved 
in water slightly acidified with hydrochloric acid. 

Sulphides, Pyrites Ore, etc. The ore should be oxidized with bromine or 
by roasting, previous to the acid treatment. 

SEPARATIONS 

Removal of Silica. The solution obtained by one of the above procedures 
is evaporated to dryness and the silica dehydrated at 110° C. for an hour. The 
residue is now extracted with dilute hydrochloric acid. Silica remains insoluble 
and may be filtered off. The solution contaia lime, together with iron, alumina, 
magnesia, etc., as chlorides. 

Removal of Iron and Alumina. The filtrate from the silica residue is 
treated with a few drops of nitric acid and boiled to oxidize the iron. Ammonia 
is now added cautiously until the solution just smells of it (a large excess over 
that required to neutralize the acid and combine with iron and alumina, 
will tend to dissolve Al(OH)j). The precipitated hydroxides are allowed to 
settle and then filtered hot through a rapid filter and washed with hot water. 
Calcium, together with magnesium, is in solution and passes into the filtrate. 

Removal of Copper, Nickel, Cobalt, Manganese, Zinc, and Elements 
Precipitated as Sulphides in Acid and Alkaline Solutions. This separation 
is required seldom in lime-bearing ores. In analysis of pyrites and certain 
other ores, containing members of the hydrogen sulphide and ammonium sul- 
phide groups, the removal of these impurities is necessary. 

The solution from the residue of silica is made slightljr ammoniacal and 
HyS passed into the solution to saturation (or ammonium sulphide may be 



90 CALCIUM 

added). The precipitated sulphides are filtered off from the solution heated to 
boiling. The filtrate containing the calcium is boiled down to 50 to 75 cc. and 
the precipitated sulphur removed by filtration. Calcium is determined in the 
filtrate by precipitation with ammonium oxalate or oxalic acid according to 
directions given later. 

Separation of Calcium from Magnesium and the Alkalies. In the pres- 
ence of considerable amounts of calcium and comparatively small quantities 
of magnesium the oxalate method of precipitating calcium, in presence of 
ammonium chloride, is generally sufficient for precipitating calcium free from 
magnesium and the alkalies. In analysis of dolomite, MgCOs'CaCOs, and of 
samples containing comparatively large amounts of magnesium, a double pre- 
cipitation of calcium is generally necessary for removal of occluded magnesium. 

Separation of Calcium from Barium and from Strontium. The alkaline 
earths are converted to nitrates, all moisture expelled by heat, and calcium 
nitrate extracted from the insoluble nitrates of barium and strontium by a 
mixture of anhydrous ether and absolute alcohol, in equal parts, or by boiling 
the dry nitrates in amyl alcohol (6.p., 137.8° C). Details of the procedure are 
given under Separations of the Alkaline Earths in the chapter on Barimn, page 53. 

Phosphate Rocks, Calcium Phosphate, etc.^ 

Determination of Lime in Presence of Phosphates, Iron, and Alumina. 

Should phosphoric acid be present in the solution, calcium will be precipitated 
as a phosphate upon making the solution neutral or slightly alkaline with 
ammonia, and will remain with iron and alumina precipitates. 

Precipitation of Calcium Oxalate in Presence of Iron and Alumina. 
The solution containing the phosphates freed from silica is oxidized by 
boiling with nitric acid as usual. Ammonia water is added to the cooled 
solution until a slight precipitate forms, and then citric acid is added in suf- 
ficient quantity to just dissolve the precipitate. If this does not readily occur, 
additional ammonia is added, followed by citric acid until the solution clears, 
then about 15 cc. of citric acid in excess. The solution is diluted to 20Q cc. 
and heated to boiling. Calcium oxalate is now precipitated by addition of 
ammonium oxalate. Iron and alumina remain in solution. 

Citric acid is made by dissolving 70 grams of the acid, HjC^HsOt • HiO, in a 
liter of water. 

Wagner's Solution. In place of citric acid, the following solution may be 
used. Twenty-five grams of citric acid and 1 gram of saUcylic acid are dissolved 
in water and made to 1000 cc. Twenty-five to 50 cc. of this reagent is effective 
in preventing precipitation of iron and alumina. 

* Zeit. flir Angewandte Chemie, 34, 776, Aug., 1898. 



CALCIUM 91 

GRAVIMETRIC DETERMINATION OF CALCIUM 
Precipitation of Calcium Oxalate and Ignition to Calcium Oxide 

Calcium oxalate is precipitated from feebly ammoniacal solutions or from 
solutions acidified with acetic, oxalic, citric, or salicylic acids, by means of 
ammonium oxalate. The presence of ammonium chloride hinders precipitation 
of magnesium and does not interfere with that of calcium. If, however, much 
magnesium (or sodium) is present it will contaminate the calcium precipitate 
so that a second precipitation is necessary to obtain a pure product. The 
compound formed from hot solutions is crystalline or granular and filters readily, 
whereas the flocculent precipitate formed in cold solutions does not. Calcium 
oxalate, CaC204-HjO,^ decomposes at red heat to CaO, in which form it is 
weighed. 

Procedure. If the calcium determined is in the filtrate from previous 
groups, hydrogen sulphide is expelled by boiling and the precipitated sulphur 
filtered off, the solution having been concentrated to about 200 cc. The fil- 
trate should contain sufficient anmionium chloride to hold magnesium in solu- 
tion in presence of ammonium oxalate (i.e, about 10 grams NH4CI per 0.0015 
gram MgO per 100 cc. of solution.) ' If not already present, the chloride is 
added in sufficient amount, and the solution diluted to about 400 cc. 

Precipitation. The solution is heated to boiling and 10 cc. of acetic acid 
added to the neutral mixture. Fifteen cc. or more of a saturated solution of 
oxalic acid * is added and after five minutes a slight excess of ammonia. The 
solution is allowed to cool an hour or so, the clear solution decanted through 
a 10-cm. filter and the precipitate washed three times by decantation and finally 
on the filter with dilute anunonia (1 : 10), or 1% ammonium oxalate. 

To remove clinging impurities (Na or Mg) the precipitate is dissolved in 
dilute nitric acid (1 : 4) and the filtrate collected in the beaker in which the 
first precipitation was made. The solution is heated to boiling after addition 
of a few drops of oxahc acid and sufficient ammonium hydroxide to make the 
solution slightly alkaline. The precipitated oxalate is allowed to settle, filtered 
and washed as in the first precipitation, the oxalate adhering to the sides 
of the beaker being carefully "copped" out. The oxalate is ignited wet in a 
weighed crucible, the heat being low at first, until the filter has charred and 
then to the full heat of the M^ker blast lamp. Fifteen minutes of blasting 
should be sufficient to obtain constant weight. If the precipitate is large a second 
igntion is advisable to insurre the complete decomposition of the oxalate and 
carbonate to oxide. 

The crucible is cooled in a desiccator and weighed as soon as possible. < 

Factors. CaO X 0.7146 =Ca, or X 1.7847 =CaCO,, or X 2.8908 =Ca(HC03)t, 
or X 2.428 =CaS04. 

^Calcium oxalate dried at 100 =CaCi04 -1120. Heated to 200** C.= =CaC204. 
At 500** C. the oxalate begins to decompose, free carbon is Uberated, and calcium 
carbonate begins to form. At bright red heat carbon burns off and the carbonate is 
completely decomposed to the oxide and COj. 

* Mellor, **A Treatise on the Ceramic Industries,;' 213 (1913). 

' Approximately 8.6% at 20** C. About five times as much ammonium oxalate 
as is required for combination with calcium and magnesium should be added to the 
solution. 

* Calcium oxide absorbs moisture and COs from the air. 



92 CALCIUM 

Other Methods. Qravimetric 

Calcium may be converted to carbonate, sulphate or fluoride and so weighed. 
The oxide above obtained may be converted to sulphate by mdstening with 
a few drops of water and then adding a slight excess of sulphuric acid (1:4, 
dilute). The excess sulphuric acid is driven off by heating over a low flame 
to SOj fmnes and then more strongly at dull red heat until the excess acid has 
been expelled. A ring burner reduces the risk of spurting. Addition of a drop 
or so of ammonia to the cooled residue and reheating assists expulsion of the 
acid. The residue is weighed as CaSOi. 

CaSO4X0.2943=Ca or X0.4119=CaO or X 0.7352 =CaCO,. 

VOLUMETRIC DETERMINATION OF CALCIUM 
Titration of the Oxalate with Permanganate ^ 

This procedure may be applied successfully in a great variety of instances 
on account of the readiness with which calcium oxalate may be separated. 
In the presence of iron, alumina, manganese, magnesia, etc., it is advisable to 
make a reprecipitation of calcium oxalate to free it from adhering contaminations. 

The following reaction takes place when potassium permanganate is added 
to calcium oxalate in acid solution: 

5CaC,04+2KMn04+8Hi^04 = 5CaS04+Ki^04+2MnS04+ 10COa+8H,O. " 

Procedure. Calcium oxalate, obtained pure, by precipitation and washing 
according to directions given under the gravimetric determination of calcium, 
is washed into a flask through a perforation made in the filter paper, the -filter 
is treated with a little warm, dilute sulphuric acid ^ and the adhering oxalate 
dissolved and washed into the flask. About 25 cc. of dilute sulphuric acid, 
1 : 1, is added and the solution diluted to 250 to 300 cc. 

When the precipitate has dissolved, the solution warmed to 60 or 70** C. 
is titrated with standard potassium permanganate, added cautiously from a 
burette with constant agitation, until a faint permanent pink color is obtained. 

One cc. N/10 KMnO* =0.0020 gram Ca,» or X0.0028 =CaO. 

Factors. CaX 1.3993 =CaO or X2.4974=CaCO, or X3.3975=CaS04 or 
X 2.581 =Ca,(P04)2. 

Analysis of Limestone and Cement. See chapter on Cement by Richard K. 
Meade. 

^ Fresenius, Hempel, Mohr, Sutton and others have testified to the accuracy 
of this methoa for the determination of calcium. 

* HCl in moderate quantity may be used in place of sulphuric acid without danger 
of liberating free chlorine as is the case in presence of iron. — Fleischer. 

* From the reaction 2KMn04, equivalent to 50 or lOH, reacts with 5CaCj04. 
and 5Ca=(5X40)-i-10 = 20. A normal solution of calcium =20 grams Caper liter. 
One cc. N/10 solution =0.002 gram Ca. 



CARBON 

Wilfred W. Scott. 

Cf^ at.wt. 12.0; 8p.gr, amorp. 1.75-2.10; cryst,: graphite, 2.25; diamond, 
3.47-3.5585; m.p. sublimes at 3500"" C; oxides, CO and COs 



DETECTION 2 

Element. Carbon is recognized by its appearance and by its inertness 
towards general reagents. It is seen in the charring of organic matter when 
heated or when acted upon by hot concentrated sulphuric acid. 

Upon combustion with oxygen or by oxidation with chromic and sulphuric 
acids, carbon dioxide is formed. The gas passed into lime water forms a white 
precipitate, CaCO^. White precipitates are formed when the gas is led into 
baryta water (BaCO« ppt.), or into an ammoniacal solution of lead acetate 
(PbCO, pptd.). 

Carbon Dioxide. Carbonates. COs in Gas. A white precipitate with lime 
water, baryta water, ammoniacal solutions of calcium, or barium chlorides, or 
lead acetate (basic). 

Carbonates. Action of mineral acids cause effervescence, COs being evolved. 
The gas is odorless (distinction from SOs, HsS, and NsOj) and is 
colorless (distinction from NsOa). The gas absorbed in the 
reagents above mentioned produces a white precipitate. The 
test is best made by placing the powdered material in a large 
test-tube with a stopper carrying a funnel and delivery tube as 
shown in the illustration. Fig. 9. For small amounts of combined 
COs, warming of the test-tube may be necessary. Sulphuric or 
phosphoric acid should be used to liberate the gas, which is 
conducted into the reagent used for the test. 

Distinction between Soluble Carbonates and Bicarbonates. 
The solution of the former is alkaline to phenolphthalein indicator 
(pink). Bicarbonate solutions remain colorless withthis indi- 
cator. Normal carbonates precipitate magnesium carbonate when 
added to magnesium sulphate solution; bicarbonates cause no 
precipitation. 

Free Carbonic Acid in Water in Presence of Bicarbon- 
ates. 0.5 cc. of rosolic acid (1 part acid in 500 parts of 80% 
alcohol), produces a red color with bicarbonates in absence of free COs, and a 
colorless or faintly yellow solution when free COs is present. 

Carbon Monoxide. The gas bums with a pale blue flame and is not ab- 
sorbed by potassium hydroxide or lime water (distinction from COs). It is oxi- 
dized to COs and so detected. With hot, concentrated potassium hydroxide, 
potassium formate is produced. 

The gas is detected in the blood by means of the absorption spectrum. 

^ Van Nostrand's Chemical Amiual, Olsen. 
' Prescott and Johnson, "Qual. Chem. Anal." 

93 




M'j COj 



Fig. 9.— Test for 
Carbonate. 



ESTIMATION 

The dement occurs free in nature in the crystalline forms, diamond and 
graphite, and in the amorphous form, charcoal, coke, etc. It occurs in iron, 
steel, and in certain alloys. Its estimation in these metals is generally required. 
Carbon is determined in the analysis of organic compounds in which it is 
invariably combined and may also be present as free carbon (asphaltum). 

Combined as a carbonate it occurs in a large number of substances, among 
which are found calcite, marble, limestone, dolomite, magnesite, strontianite, 
witherite, spatic iron ore. It occurs as the dioxide in the air, in water (HiO -COi) 
and in flue gas. Carbon dioxide is the active constituent of baking powders 
(NaHCO,). 

Preparation of the Sample 

Iron, Steel, and AUoys. Drillings taken from different sections of the rep- 
resentative bar should be free from grease and dust. These are best kept in 
glass-stoppered bottles. Where a large number of 
daily samples are determined, it is found more 
convenient to use small manila envelopes, upon 
which the record of the analysis may be placed. 
Should it he impossible to obtain drillings free 
from grease, this impurity may be removed by 
heating the sample in an atmosphere of nitrogen, or 
by repeated extraction with ether. 

Coarse chips, cast-iron driltii^, etc., should be 
broken down in a chilled-steel mortar, Fig. 10. 
Fig. 10— Chilled Steel Mortar. Carbon may now be separated in a definite 
weight of the sample as directed below, or it may 
be determined by direct combustion or by oxidation with chromic acid accord- 
ing to a procedure given later. 

Organic Matter. It is advisable to fuse this in a nickel or iron crucible 
with sodium peroxide. The carbonate thus formed may be determined as usual. 
The organic substance may be oxidized directly in the combustion furnace. 

Carbonates. Limestone, Dolomite, Cement, Alkali Carbonates and 
BicarbonateB. The powdered material is decomposed by addition of an acid 
as directed in the methods given later. 

Separation of Carbon from Other Substances 

The element is generally determined as carbon dioxide, in which form it is 
liberated from most of the combinations in which it occurs, free from other 
substances by ignition in a current of oxygen, or by oxidation with chromic 
acid as directed later. 

Separation of Carbon in Iron and Steel. Cupric Potassium Chloride 
Method. 0.5 to 2 grams of the drillings arc treated with 100 to 200 cc. of 
cupric pota-'tsium chloride solution and 10 cc. of hydri^hloric acid (1.19). Tliis 
mixture dissolves the iron acconling to the reaction 

Fe-|-CuCli-FeCI,+Cu and Cu-i-CuCI, -CuiCli-|-carbon as a residue. 




CARBON 95 

The solution should be stirred frequently to hasten the solution of the iron. 
It is advisable to keep the temperature of the solution at about 50° C. When 
the iron and copper have dissolved the carbon is filtered off into a perforated 
platinum boat or crucible, as directed under the methods. It is now oxidized 
to COj and so determined. 

Note. The cupric potassium chloride solution is prepared by dissolving 150 
parts of potassimn chloride and 170 parts of ciystallized cupric chloride in water 
and crystallizing out the double salt. Three huncired grams of this salt are dissolved 
in 1000 cc. The solution may be used several times by chlorinating the dirty brown 
filtrate from the carbonaceous residue. The cuprous chloride formed during the 
solution of the steel is converted again to cupric chloride, and the chlorinated double 
salt is even more energetic in its solvent action than the freshly made reagent. (Blair.) 



GRAVIMETRIC METHODS FOR DETERMINATION OF 

CARBON 

The determination of carbon by combustion with oxygen is made in two 
general classes of substances: A. Steel, iron and in certain alloys. B. Organic 
compounds. Carbon in steel and alloys is considered in two forms: carbide 
or combined carbon, and graphitic carbon. In organic substances carbon occurs 
principally combined with hydrogen, oxygen, and nitrogen. For the present 
we will consider procedures for the determination of carbon in steel and 
alloys. 

The most accurate procedure for determination of carbon in steel, alloys, 
and in other materials containing the substance combined or free is by com- 
bustion with oxygen in a furnace heated by gas or electricity; the carbon dioxide 
formed being absorbed in caustic, and weighed. 

Apparatus. Combustion Furnace. Although the gas furnace has 
been used more commonly on account of gas being more available than elec- 
tricity, the extension of generating electric plants makes it possible to use electric 
furnaces, and these are gradually displacing those heated by gas, as they are 
more compact, easily manipulated and comparatively simple in structure. 

A simple electric furnace may be made by wrapping a silica tube with a 
thin covering of asbestos paper, which has been moistened with water. On 
drying the paper will cling to the tube. A spiral coil of nichrome wire (Driver 
and Harris) is wound around this core. On a 2-foot length of tube two 45-foot 
lengths of No. 18 wire, connected in parallel, will heat the tube to bright redness, 
attaching the terminals to an ordinary light socket. The coils should be 
covered with }-in. coating of alundum cement. For appearance' sake as well 
as for protection, the tube is placed in a large cyhnder of sheet iron, packed 
around with asbestos, and is held in position by circular asbestos boards placed 
at the ends of the large cylinder. The cylinder is mounted on a stand. 

Absorption Apparatus. A large number of forms are for sale. The Geissler 
and Liebig bulbs have been popular (Figs. 11 and 12), but are now being displaced 
by forms that have less surface exposed, that are more easily cleaned and less 
fragile, such as Gerhardt's, Vanier's and Fleming's apparatus (Figs. 13, 
14 and 16). The Vanier and the Fleming absorption apparatus are especially 
to be recommended, on account of their capacity, compactness, efficiency, in 
handling gases passing at a rapid rate, and their simphcity of form. 



96 CARBON 

Procedure for Determining Carbon by Combustion. Mr. William R. 
Fleming ' describes his apparatus in the Iron Age, Jan. 1, 1914. The following 
abstract is taken from Elmer & Amend's circular, edited by Mr. Fleming. 

The greatest value of this rapid method is realized when it is used to follow 
a bath of steel in the open-hearth furnace preliminary to tapping. It abolishes 






Fio. 11.— Geiflsler Bulb. Fia. 12.— IJeblg Bulb. Fia. 13.— Gerhardt Buib. 

completely the unreliable and dangerous color carbon. By this method abso- 
lutely accurate results can be reported to the open hearth ten minutes after 
the drillings are received. 

In principle this method is not new; In manipulation it ia new. Hereto- 
fore chemifts have been laboring under the impression 
that the flow of gas during a combustion must noC^xceed 
a certain snail-like pace. This false impression has been 
/^_^ r"-— -» injected into the minds of chemists by a few who were 

1 1 o;^ iT""^ supposed to have investigated the matter. The truth is 

/ yQ: ^^^^ ^^^ faster oxj-gen is fed to burning steel the more 

/ \ J"'^Z::^x complete the combustion will be. The rate of current is 

' ' ^ ' ' limited by the efficiency of the apparatus used to absorb 

the evolved carbon dioxide. 

The Apparatus Described. The combustion train is 
shown complete in Fig. 15. The oxygen is delivered to 
the train ttmiugh a regulating and reducing valve such as 
is used for welding. The regulating valve is not essen- 
tial, yet any chemist who uses one will appreciate its con- 
venience, especially in this method. Ita convenience will be 
explained later. /C is a mercury pres-sure gauge. It serves 
Bottle. aa a guide during the combustion and is an essential piece 

of apparatus. The graduated column is 6 ins. high and is 
divided into eighths. P is a washing bottle containing caustic potash solution. 
Filled to the mark indicated with 50% solution it will ser\-e for at least 1000 com- 
bustions. It is used solely to indicate the flow of gas, not to purify it. If the 
chemist desires he may omit this from the train. T is a calcium chloride jar. 
It is filled to the mark indicated with finely divided calcium chloride, alwut pea size, 
retaining all the dust. \ layer of asbcst^is is fonncd over the chloride and the 
remaining space filled with soda lime. The glass tubing leading from the jar is 
loosely packed for a distance of several in<!hes with asbestos. This prevents any 
soda lime dust being carried into the combustion tube. (? Ls a mercury valve like 
that used in Johnson's train. It is used solely to maintain an atmosphere of 
' Metallurgist, Andrews Steel Company, Newport, Ky. 




CARBON 



97 



pure oxygen in the purifying train, a condition essential to accurate results. 
It is not used to prevent carbon dioxide backing into the purifying train, of which 
there is not the remotest possibility. 

The combustion tube is the ordinary fused n a,- * P 

silica tube glazed on the inside onlj'. The tube t W / . '^ r 

is 30 ins. long with inside diameter of from J f^ — ''^^ 

to I in. One tube of 30 ins. will serve twice as 
long as one of 24 ins. It is loosely packed with 
asbestos for a distance of 6 ins. at the exit end, 
and 3 ins. is allowed to project from the furnace. 
For about the first 100 combustions, the com- 
bustion boat is pushed close against the asbestos. 
The portion of the tube immediately above this 
will become coated with iron oxide. The asliestos 
is then moved up so that it covers this portion 
of the tube and a fresh area exposed to the 
spraying oxide. In this manner one tube can be 
made to serve 600 combustions or even more. 
Both platinum and nickel cylinders have been 
used inside the tube to protect it from the spraying 
oxide, but it is doubtful whether this practice pays. 
These cylinders are not used in this laboratory 
because it ia believed that they delay incipient 
combustion for at least thirty seconds. 

The Furnace and Combustioa Tube. The 
furnace used is one of the ordinary resistance 
type. It is constantly maintained at a tem- 
perature of 1000° C. "This temperature is verified 
daily by the use of a pyrometer. Many claim to 
be expert at judging temperatures, but none are 
expert enough to be without a pjTometcr. The 
two-way stop following the combustion tube mil 
be found very convenient when it is not desirable 
to pass the current through the jars Z and O. 

Z is filled with 20-mesh zinc. Once filled it 
will serve for several thousand combustions. As 
a matter of fact it is included in this train as ., ^ ^ 
a filter. If nickel boats and aluminum are used ^^ $ 
the chemist may omit this zinc jar from the ^ 

train, for with all ordinary grades of steel it serves 
no purpose. 

is the phosphoric anhj-dride jar. A little 
asbestos is placed in the lower part just above 
the lower stopper. The remaining space in the 
jar is completely filled with phosphoric anhydride. 
The upper stopper is packed tightly enough to 
prevent any powder being swept 'into the weigh- 
ing apparatus. As the anhydride liquefies it iC^^ 
passes down into the lower stopper, where it can 
be removed conveniently without disturbing the anhydride above it. Likewise 




a 



I 



^ 
^ 



.s><=- 




Soda 
Lime 




98 CARBON 

the anhydride can be replenished by removing only the upper stopper. The 
jar need not be washed oftener than once in 500 combustions. When filled 
with anhydride, fresh reagent need not be added for at least 150 combustions. 
After each combustion the jar should be given a few sharp taps with the hand 
to prevent canals being formed. 

Details of the Absorption Apparatus. The absorption apparatus, shown 

in detail in Fig. 16, has been modified slightly at the sug- 
gestion of Henry G. Martin, of the Railway Steel Spring 
Company, Chicago Heights, 111. This apparatus is no more 
• efficient than the old style, but it is much more convenient 
and less troublesome. In the old-style tube the anhydride 
would liquefy after several days and require replenishing. 
To overcome this, Mr. Martin suggested using separate cham- 
bers for the anhydride and soda lime, so that conmiunication 
could be broken when the tube was standing idle. The tube 
shown. Fig. 16, is Fleming's modification of Mr. Martin's 
suggestion. When properly filled this tube will serve for at 
least 70 combustions when operating on 1.5 grams of sample 
containing 1.03% carbon. 

The anhydride in the upper chamber serves for at least 
300 combustions. Soda lime, placed in the lower tube in alter- 

„ -J- . nate layers (J in.) of the different meshes, has proven a very 

Absorption"^^ convenient and desirable reagent. The 12-mesh soda lime for 
paratus. nitrogen can also be used with excellent results. If this is 

employed, part of it should be ground to about 60-mesh and 
alternate layers of fine and coarse used. 

It is exceedingly important that the tube be loaded with alternate layers 
of coarse and fine reagent, for, if the 12-mesh reagent is transferred directly 
from the bottle to the absorption tube, the latter will fail to be effective for 
more than 30 combustions and in some cases less. The reason for this is evident. 
The lower stopper is packed loosely with asbestos, also the lower portion of the 
soda lune chamber just above the stopper. Beginning with a layer of 
12-mesh soda lime, the entire chamber is filled with alternate layers of fine and 
coarse reagent. The small diameter of the anhydride chamber is packed with 
asbestos and the remaining space filled with phosphoric anhydride. Finally, 
the upper stopper is packed with asbestos. The anhydride chamber, filled as 
indicated, will not require refilling for at least 300 combustions. It is not 
necessary to turn the chamber to break conununication while the tubes are 
idle, for the packing of the small diameter with asbestos prevents the absorption 
of moisture from the soda lime. The tubes must be used in pairs, so that one 
serves as. a tare in weighing the other. A pair of tubes assures the operator 
of at least 140 combustions. A glass or rubber tubing about 12 ins. long serves 
as a guard for the absorption tube. It connects the bottle, Pi, which is used 
to indicate flow of gas. 

The use of clay boats has been abandoned in favor of nickel boats filled with 
alundum. These are greatly superior to clay boats in every conceivable way. 
The alundum is labeled as being free from carbon, but this is not true. In fact, 
some of it contains considerable carbon. It should always be burned in oxygen 
at 1000® before using. The boats are formed out of 22-gauge pure sheet nickel. 
One boat will serve for about 100 to 150 combustions, some more, some less. 



CARBON 99 

Details of the Analysis. The furnace being at 1000^, the two freshly 
prepared absorption tubes are placed in the train and oxygen run through at 
the rate of 300 cc. per minute for fifteen minutes. This insures the displace- 
ment of all air from the purifying train as well as the absorption tubes. Remove 
one absorption tube from the train and turn on the oxygen until the mercury 
stands at about 2 ins. The rate of current is then measured by inverting a 
graduated cylinder filled with water. Several trials will establish a rate of 
about 325 cc. per minute. Note the reading of the column of mercury at this 
rate and subsequently, when using the same absorption tube, maintain this 
same pressure in the train and the rate of flow will be 325 cc, the rate during all 
combustions. Shut off the oxygen and, when it comes to a slow bubbling 
through Pi, close the upper stopper of the absorption tube. Disconnect it from 
the train, but do not close the lower stopper for about five seconds after dis- 
connection. Weigh against its mate as a tare. It is now ready for the first 
combustion. 

Weigh 1.5 grams of drillings, preferably thin, curly drillings from a twist 
drill, and spread out in the nickel boat which is half filled with alundum. Place 
the absorption tube in the train and place its mate beside it. With the oxygen 
flowing about 100 cc. per minute, the drillings are pushed into the combustion 
tube. The current is immediately run up to the desired pressure, which gives 
325 cc. per nunute. The regulator will do the rest. It will feed the oxygen 
automatically to the burning steel. As a rule the drillings are entirely burned 
one and one-half minutes after insertion. Continue the flow of oxygen for 
three and one-half minutes more (five minutes, total time) and disconnect 
as before the absorption tube. Weigh immediately. The result will be accurate 
and reliable. Whether determining carbon in a standard steel, where the 
greatest accuracy is required, or in a bath test, the time required is always 
five minutes. 

The weight of the boat, plus refractory lining, should be kept as low as 
possible, so as not to introduce too much cold material into the combustion 
tube. The boats used in tWs laboratory are i in. wide, i in. deep and 3 ins. long. 
Sheet nickel varies in percentage of carbon. As a rule, a nickel boat must be 
ignited in oxygen at 1000® for one to two hours. 

There seems to be a difference of opinion concerning the physical condition 
of the steel after burning, some chemists believing that inaccurate results are 
obtained if the drillings have fused during combustion. Others maintain that 
complete fusion of the drillings is essential to accurate result. If drillings 
which happen to be a little thick are used, low results are obtained unless these 
are perfectly fused. 

Graphitic Carbon 

In Iron and Steel. The sample of 1 gram of pig iron or 10 grams of 
steel is treated with 15 cc. of nitric acid (sp.gr. 1.2), per gram of sample taken. 
When all the iron has dissolved, the graphite is allowed to settle and the super- 
natant liquid decanted onto an ignited asbestos filter, using either a perforated 
boat. Fig. 17, or a filtering tube. The residue is transferred to the filter, 
and washed thoroughly with hot water. It is treated with hot caustic 
solution (sp.gr., 1.1), washed thoroughly again with hot water, then with a 
little dilute hydrochloric acid, and finally with hot water. The carbon is 
now burned by one of the procedures given — the oxidation in the combustion 



100 



CARBON 



furnace being recommended. The CO2 is absorbed in caustic and estimated 
according to the standard procedure given for carbon. 

002X0.2727= graphitic carbon. 

The perforated boat, shown in the cut, fits snugly 
into the receptacle below. Sufficient asbestos is poured 
into the boat to form a film over the bottom. A seal 
is made around the boat with additional asbestos, the 
apparatus having been inserted in a rubber stopper 
in the neck of a suction flask and suction applied. 

The apparatus is recommended by Blair for com- 
bustion of graphitic carbon or of total carbon liberated 

•n. -i^ T» X JTT ij from iron or steel by the cupric potassium chloride 
Fig. 17. — Boat and Holder ... rr.i i i. u 1 j j- 4.1 : ^.u^ 

for Carbon Determina- "method. The boat may be placed directly m the 

tion. combustion tube and the carbon oxidized as usual. 




m r ' 


6 

1 






The Shimer Combustion Apparatus ^ 

The apparatus, designed for the rapid determination of iron and steel, is 
in general applicable to the same class of chemical operations as is the com- 
bustion tube of platinum, silica, or porcelain. It 
offers the advantage of neatness, reduction in the 
number of parts to be handled, diminished con- 
smnption of gas, and increased ease of manipula- 
tion. The simplified form, shown in the cut. Fig. 
18, enables the use of the standard form of plati- 
mnn crucible, A, with its inner wall ground to fit a 
tapered nickel, water-jacketed stopper, B. The rub- 
ber jacket of the original type is eliminated and a 
detachable nickel reinforcing ring, C, at the top of 
the crucible serves the double purpose of completing 
the security of the seal and as a support for the 
apparatus. B O 

Water is circulated through the stopper through Fig. 18.— Shimer Combustion 
the tubes c and d. The current of oxygen passes Apparatus, Simplified Form. 
through a into the crucible, oxidizing the material 

within the crucible, which is heated to red heat with a burner placed below it. 
The carbon dioxide formed passes through b to the absorption train. The 
remainder of the apparatus for the determination is the same as is used with 
the combustion tube. An asbestos shield protects the upper portion of the 
outfit, the crucible fitting snugly in a hole in the asbestos board. 



^^m 



Combined Carbon 

Indirect Method. The excess of carbon remaining when the graphitic 
carbon is subtracted from total carbon (in iron and steel), is calculated as com- 
bined carb(m. This difference method is generally accepted as being the most 
accurate for estimation of combined carbon. 



* Courtesy of Baker Platinum Works. 



CARBON 101 

Notes. In cluromiuin, tungsten and titanium steels a temperature of 1500*' C. 
is necessary to oxidize the carbon by direct combustion for thu-ty minutes. (J. R. 
Cain and H. E. Cleaves, J. Wash. Acad. Sci., 194, 4, 39a-397.) 

Carbon in Soils. One to 3 grams of 60^mesh sample is treated with a solu- 
tion of 3.3 grams CrOi+10 cc. H^O and 50 cc. cone. HSO4 (1.84). The evolved 
CO2 is absorbed in standard caustic and titrated with acid, phenolphthalein and methyl 
orange being used as indicators. (J. Ind. Eng. Chem., 1914, 6, 843-846.) 

DETERMINATION OF CARBON IN ORGANIC SUBSTANCES 

Combustion of Organic Substances Free of Nitrogen, Halogens, 

Sulphur, and the Metals 

The following modification of the procedure described for determination 
of carbon in iron and steel is applicable to the determination of carbon in organic 
substances free from the substances mentioned above. 

Apparatus. This is practically the same as that shown in Fig. 19, with the 
exceptions that copper plugs may be used to advantage in place of the plati- 
num plugs. In the absorption end of the train a calcium chloride tube is 
preferred. The calcium chloride should have been saturated with dry COi 
gas, the excess of which has been removed by a current of pure air. This tube 

Copper 
Flatinum Boat. Flatinurti Oxide /Platinum 



=^ cA^ -^_'_JKd^^ 



-Qgag 




Fig. 19. — ^Diagranmiatic Sketch of Combustion Tube. 

is weighed as well as the potash bulb, the calcium chloride retaining the water 
formed by the combustion of the hydrogen of the organic substance, which 
is thus determined. 

The organic substance, if a solid, is introduced into the combustion boat 
directly; if it is a liquid, it is held in a bulb blown in a capillary tube. One 
end of the tube is sealed and a bulb blown. When cool, the tube is weighed, and 
the material then introduced by first warming the bulb and then inserting the 
open end of the tube into the liquid to be examined. By cooling the bulb, liquid 
is drawn into the tube. The end is wiped off, and the liquid expelled from 
the capillary by gently heating this portion. The end is now sealed if the 
liquid is volatile, otherwise it is left open, and the tube is weighed. The 
increased weight is due to the organic substance. The tip of the capillary is 
now broken, if sealing was necessary, by means of a file. The tube containing 
the sample is placed in the boat, the open end of the capillary pointing toward 
the open end of the combustion tube. After connecting up the apparatus, 
the copper oxide end of the tube is heated to redness and oxygen slowly passed 
through the tube at such a rate that the bubbles in the potash bulb can be 
readily counted. The entire tube is now heated and remaining operation is 
the same as has been described for iron and steel combustion. 

The gain of weight of the calcium chloride tube is due to water formed by 
the combustion of hydrogen of the compound, that of the potash bulb to the 
carbon. 

H2OXO.II2I =H, 
CO2X 0.2727=0. 



102 CARBON 

Note. The oxygen gas should be free from hydrogen. A preheater, placed 
before the purifying tubes of the train, causes the combustion of the hydrogen and 
the absorption of the water formed before the gas enters the combustion tube. 

Determination of Carbon and Hydrogen in Nitrogenous 

Substances 

A modification of the first procedure described for determinations of carbon 
and hydrogen in organic substances must be made, since from substances 
containing nitrogen, nitroso and nitre compounds, oxides of nitrogen are formed 
which would be absorbed in the calcium chloride and potash bulbs, giving high 
results for hydrogen and carbon. To overcome this difficulty, a copper spiral, 
that has been reduced (See note below) is placed in the front end of the 
combustion tube (to the right in Fig. 19) to reduce the oxides of nitrogen to 
nitrogen. 

Note. Reduction of copper spiral may be accomplished as follows: The copper 
spiral is prepared bv rolling together a piece of copper gauze about 10 centimeters 
wide, making it as large as will conveniently pass into the combustion tube. The 
spiral is heated till it glows by holding it in a large gas flame, and while still 
hot it is dropped into a test-tube containing 1 or 2 cc. of methyl alcohol or ether. 
This quickly boils away, igniting at the end of the tube. The copper is completely 
reduced to bright metallic copper. The spiral is taken out with a pair of crucible 
tongs and dried by quickly passing it through a flame a few times, and while it is 
still warm it is introduced into the front of the combustion tube.^ 

The substance is introduced into the tube and the connections made. The 
copper oxide spiral, that was pushed after the boat, is heated, and then the 
reduced spiral (right end of tube). The oxide near the boat, and finally 
the entire tube is heated to a red heat. When the bubbles cease to show in 
the potash bulb, the stopcock is opened to the oxygen-purifying train and a 
slow flow of ox>'gen turned on, the gas allowed to pass through the tube until 
it can be detected with a glowing splinter at the exit of the absorption end of 
the apparatus. 

If the substance is difficult to burn, it is mixed with freshly ignited (cold) 
copper oxide, which assists combustion. 

The remainder of the operation is the same as has been described. 

Organic Substances Containing Halogens 

The procedure is the same as that described for nitrogenous substances 
with the exception that a silver spiral is used in place of the reduced copper 
spiral. The heating of this spiral* should be between 180 and 200° C. (not 
over 200°). 

Organic Substances Containing Sulphur 

These are best ignited vnih sodium peroxide and the carbonate formed is 
detenuined by the procedure given for carbon dioxide in carbonates. 

The Wet Combustion Process for Determination of Carbon 

The method depends upon the oxidation of carbon to carbon dioxide when 
the powdered material is digested with a mixture of concentrated sulphuric 
acid and chromic acid, or potas.sium dichromate, or permanganate. The pro- 

* Treadwell and Hall, "Quantitative Analysis." John Wiley & Sons. 



CARBON 103 

cedure is applicable to oxidation of free carbon, carbon combined in organic 
substances and in certain instances to carbon combined with metals, where 
the substance may be decomposed by the action of the acids. ^ It is of value 
in determination of carbonates in presence of sulphides, sulphites, thiosulphates 
and nitrites, which would vitiate results were they not oxidized to more stable 
forms, from passing into the potash bulb with the carbon dioxide. 

Apparatus. The apparatus is identical with that used for determining 
carbon dioxide in carbonates, Fig. 20, with the exception that in place of the 
acid bulb nearest the decomposition flask two bulbs are placed. The first of 
these contains a strong solution of chromic and sulphuric acids, the second 
is filled with glass beads moistened with chromic acid solution. Following this 
is the drying bulb containing concentrated sulphuric acid and finally the 
absorption apparatus, as shown in the illustration. 

Procedure. 0.2 to 1 gram of the powdered material, fine drillings, free 
carbon, or organic substance is placed in the decomposition flask. If the 
material is apt to pack it is advisable to mix with it pure ignited sea-sand to 
prevent this. Five to 10 grams of granular potassium dichromate are added 
and the apparatus swept free of carbon dioxide by passing purified air through 
it before attaching the absorption apparatus. The potash bulb is now weighed, 
using a counterbalance bulb and following the precautions given in the dry- 
combustion method. The bulb is attached to the train. 

Oxidation. Concentrated sulphuric acid placed in the acid funnel, attached 
to the decomposition flask, is allowed to flow down on the sample until the 
funnel is almost empty; the stop-cock is then closed. A flame is placed under 
the flask, when the vigorous action has ceased, and the material gently heated 
until the reaction is complete and the organic matter or carbon completely 
oxidized. 

The apparatus is now swept free of residual COj by applying suction, the gas 
being completely absorbed by the potash, or the soda lime reagent. 

The increase of weight of the absorption bulb is due to carbon dioxide. 

CO2X0.2727=C. 

Note. The foUowing additional purifiers are frequently advisable: (a) an absorp- 
tion bulb containing silver sulphate to absorb chlorine and vapors from sulphur 
compounds; (b) a capillary tube of silica or platinmn heated to a dull redness to oxidize 
any nydrocarbons, carbon monoxide, etc., that may be evolved and imperfectly oxidized 
by the chromic acid. 

DETERMINATION OF CARBON DIOXIDE IN CARBONATES 

The method is applicable for determination of carbon dioxide in limestone, 
dolomite, magnesite, strontianite, witherite, spatic iron ore, carbonates of sodium, 
and potassium, bicarbonates in baking powder, carbon in materials readily 
oxidized to Cd chromic sulphuric acid mixture. The procedure depends upon 
the evolution of carbon dioxide by a less volatile acid, or the oxidation of carbon. 
The CO2 is absorbed in caustic and weighed. 

Apparatus. The illustration shows the apparatus found suitable for this 
determination. It is Knorr's apparatus slightly modified. The absorption 
bulb or bottle should be one that will effectively absorb carbon dioxide entering 

^ Not applicable for determining carbon in ferro-silicon, ferro-chrome or tungsten 
steel 



104 



CAEBON 



at a rapid rate. ITie Vanier or the Flenung forms is satisfactory for this 
purpose. 

Procedure. A sample weighing 0.5 to 2 grams, according to the carbon 
dioxide content, is placed in the dry decomposition flask (C). The flaak is 
closed by inserting the funnel tube (B) fitted with the soda lime tube (A), and 
connected by meuns of a condenser to the train for removing impurities from 
carbon dioxide, leading to the absorption bulb, as shown in Fig. 20. 




Fi«. 20. — Appomtus fur DclernuniiiK C'arliini Dioxiile. 

The apparatus is swept out wilh a current of dry, purified iiir before attaching 
the weighed absorption ln)ttlo. Tliis in jiccoinplislicd by applying gtinllo suction 
at the end of the purifying train. The ulisoriiliiin ap|mr;i(us is now attached 
{Fleming absorption appariilus is sbow-n in tb<! il lust radon). The tube (B) 
is nearly filled with dilute sulphuric acid (1 : 3), the stop-i-ock (/f) iKiing closed. 
The soda lime tube is now inserts! into place as shown in the cut, Tho acid 
in (£) is now allowed to run slowly down on the sample at a rate tliat evolves 



CARBON 105 

gas not too rapidly to be absorbed; 1 to 2 cc. of acid being retained in (B) to 
act as a seal, the stop-cock (^0 being then closed. 

When the violent action has ceased, the solution in (C) is heated to boiling 
and boiled for about three minutes. If the sample is baking powder, or con- 
tains organic matter, the decomposition flask is protected from excessive heat by 
placing a casserole of hot water under it. This prevents charring of the starch 
or organic matter, which would be apt to occur if the direct flame was used. 
Gentle suction is now applied to the absorption end of the apparatus and the 
stofHcock {B') opened, allowing the remainder of the acid to flow into the flask (C) 
and admitting a current of air, purified by passing through the soda lime in 
(A). The suction should be gentle at first, and then the speed of the flow 
increased to the full capacity of the absorption bottle. A fairly rapid current 
is preferred to the old-time procedure of bubbling the gas through the apparatus 
at a snail-like pace, but discretion should be used in avoiding a too rapid flow. 

In the analysis of baking powders, where foaming is apt to occur, the decom- 
p>osition flask should be of sufficient capacity to prevent foaming over. A small 
flask is generally to be preferred for obvious reasons. By gently heating to 
boiling during the passage of the air, steam assists in expelling any residual 
COj in the flask. When the passage of air is rapid, this boiling should be dis- 
continued. 

The increase of weight of the absorption bottle is due to the carbon dioxide 
of the sample. This procedure gives total CO2. 

Determination of Carbon Dioxide by Measuring the Gas 

Fairly accurate results may be obtained by measuring the gas evolved. A 
large cylindrical tube having a capacity of about 1100 cc. is used. The tube 
is graduated from 1000 cc. to at the upper portion of the cylinder; a space of 
about 100 cc. remains at the upper portion. A tube extending from a little 
above the graduation to the bottom of the cylinder carries out the water as 
the gas is admitted. 

To make the run, the cylinder is filled to the mark with saturated salt 
solution.* It is now connected to a condenser.' Twenty-five cc. of saturated 
salt solution are admitted to the decomposition flask, and the generated gas 
measured by the water displacement in the tube described. Calculations are 
made after reduction to standard conditions. 5.1 cc. CO2 at 0° C. and 760 
mm. weigh 0.01 gram. 

Residual Carbon Dioxide 

This is the COi remaining after baking powder has been treated with water 
and the evolved COi expelled by warming. 

The procedure recommended by the U. S. Department of Agriculture is as 
follows:* 

* H. W. Bnibacker, Jour. Ind. Eng. Chera., 1915, 7. 432. 

*The nitrometer may be used in place of the cylinder and atmospheric condi- 
tions obtained as usual. Formula for reduction to 760 millimeters and O'' C. : 



760(1+ .003670* 
* Bureau of Chem. Bulletin No. 107. 



106 



OABBON 



Weigh 2 grams of baking powder into a flask suitable for the subsequent 
determination of carbonic acid, add 20 cc. of cold water, and allow to stand 
twenty minutes. Place the flask in a metal drying cell surrounded by boiling 
water and heat, with occasional shaking, for twenty minutes. 

To complete the reaction and drive off the last traces of gas from the semi- 
solid mass, heat quickly to boiling and boil for one minute. Aspirate until the 
air in the flask is thoroughly changed, and determine the residual carbon dioxide 
by absorption, as described under total carbonic acid. 

The process described, based on the methods of McGill and Catlin, imitate, 
as far as practicable, the conditions encountered in baking, but in such a manner 
that concordant results may be readily obtained on the same sample, and com- 
parable results on different samples. 

Available Carbon Dioxide 

The residual is subtracted from the total, and the difference taken as avail- 
able CO.. 

Determination of Carbon Dioxide by Loss of Weight 

An approximate estimation of the carbon dioxide in carbonates— baking 
powders, bicarbonate of soda, limestone, etc., may be obtained by the loss of 
weight of the material when treated with a known weight of acid. 





FiQ. 21.— Schroetter'B Alkalimeter. Fio. 22.— Mohr's Alkalimeter. 

Various forms of apparatus are used for this determination. The Schroetter 
and Mohr tj-pes are shown. Figs. 21 and 22. 

About 0.5 to 1.0 gram of sample is taken and placed in the bottom of the 
flask, dilute hydrochloric and strong sulplniric acids then placed in the bulbs 
as indicated in the illustrations. The apjMiratus is weighed as it is thus charged. 
The hydrochloric acid is now allowed to flow down on the carbonate and the 
stopper closed. The evolved gas passes through the strong sulphuric acid, 
which absorbs the moisture. After the vigorous action has subsided the appa- 



CARBON 107 

ratus is placed over a low flame and the solution heated to boiling and boiled 
very gently for about three minutes. COi-free air is aspirated through the 
solution to expel the last traces of COs, by applying gentle suction at a and 
opening b, the air being purified by passing through soda lime. The apparatus 
is again weighed and the loss of weight taken as the COs of the material. 

Available COi in baking powder may be determined by substituting water 
in place of hydrochloric acid. 

VOLUMETRIC METHODS FOR THE DETERMINATION OF 

CARBON 

Total Carbon. Absorption of Carbon Dioxide in Barium Hydroxide 

The carbon dioxide evolved by oxidation of the material by dry combustion 
vnth oxygen or by oxidation with chromic sulphuric acid mixture is absorbed 
in barium hydroxide, free from carbonate, and the precipitated barium car- 
bonate titrated with standard hydrochloric acid. 

Procedure. The essential difference in this method from those already 
described under the gravimetric methods is in the fact that a perfectly clear 
saturated solution of barium hydroxide is used for absorption of the carbon 
dioxide in place of caustic potash. Considerable care must be exercised to 
prevent contaminating the reagent with carbonate. The solution is drawn 
by suction through a siphon, dipping below the surface of the reagent, into the 
absorption tube, which should be of such construction that the material may 
readily be poured out. After absorption of the COa gas, the apparatus is dis- 
connected and the excess barium hydroxide neutralized with dilute hydro- 
chloric acid (1:4) using phenolphthalein indicator. A few drops of methyl 
orange are now added and a measured excess of standard hydrochloric acid 
run in from a burette. After heating to boiling the excess acid is titrated with 
standard caustic solution, 1 cc. of which is equivalent to 1 cc. of the hydro- 
chloric acid. 

HC1X0.1646=C. 

1 cc. N/10 HClc:0.0022 gram CO2. 

Note. The method is used in the Omaha laboratory of the Union Pacific Rail- 
road. Dr. N. F. Harriman, Chief Chemist, informed the writer (W. W. Scott), that 
with care no difficulty is experienced with contamination of the barium hydroxide 
and excellent results are obtained. The Victor-Meyer bulb is used for holding the 
barium hydroxide. 



Determination of Cart)on by Measurement of the Volume of Carbon Dioxide Evolved 
by Oxidation of Carbon, or by the Decomposition of Carbonates with Acid. 

Description of the Scheibler and Dietrich Process and that of Lunge and March- 
lewski are given in Mellor's work on ** Inorganic Analvsis," pp. 555-559, 1st Ed. 
A modification of Wiborg's method is described in Blair, 'Chemical Analysis of Iron," 
pp. 146-149, 7th Ed. 

Determination of Carbon Dioxide in a Gas Mixture. 
See Gas Analysis. 



Direct Colorimetric Method for Determination of Combined 
Carbon 

The procedure is of value to the eteel laboratory where a large number of 
daily determinations of combined carbon are required. By this method over a 
hundred doterminations a day may be made by an experienced manipulator. 
The method depends upon the color produced by combined carbon diasolved 
in nitric acid, the depth of color increasing with the combined carbon content 
of the material. Comparison is mode with a standard sample of iron or ateel, 
which ia of the same kind and in the same physical condition as the material 
tested.' That is to say, a Bessemer steel should be compared with a Bessemer 
standard, open hearth with open hearth, crucible steel with crucible steel, the 
standards containing approximately the same amounts of carbon, and as nearly 
as posflible the same chemical composition. The samples should be taken 
from the original bar which has not been reheated, hammcreti, or rolled. Copper, 





Fia. 23. 
Hot Water Racks for ": 



t Tubes. 



cobalt, and chromium will interfere with the test; the other elements have veiy 
little effect. 

Procedure. One standard sample of 0.2 gram and the same amount of 
sample drillings are taken for analysis. The weighings are conveniently made 
in brass or aluminum pans, boat-shaped to enable the drillings to be dumped 
into test-tubes. A counterpoise, weighing the same as the boat, ia placed on 
the opposite pan, together with the 0.2 gram weight. A magnetized knife will 
assist in removing the excess of material. The weighed sample is brushed into 
a test-tube 6 ins. long (150 nun.) g in. (16 mm.) in diameter. (Each test-tube 
has a label near the open snd to distinguish the sample.) A rack or a 600-cc. 
beaker may be employed for holding the test-tubes during the weighing. After 
the batch is rea<ly the tubes are transferred to a perforated rack (Figs, 23 or 24) 
and this then stood in the water buth filled with cold water. 

The pro|)Cr amount of nitric acid (sp.gr. 1.2; e.g., 1 cone. HNOj : 1 HiO)i 
from a burette, is now added t^ each test-tul)e. 



3 cc. HXO, for 0.;t% C, 

4 cc. HXO, for 0.3 to O.,-)-:;.. 

5 cc. HNO. for 0.5 to 0.8% C. 



•. HXOiforO.S tol%C. 

:. HXO, for over 1% C. steel'. 



' Blair, "The Chemical Analysis of Iron." 



CARBON 



109 



The depth of color produced by the acid will give an idea of the amount 
required. One cc. of acid is added at a time until the depth of color is correct. 
This reqiures experience gained from observation of the color produced by standard 
samples. The acid is added slowly to the coarse drillings. Insufficient acid 
gives a darker tinted solution than it properly should be. The nitric acid should 
be free from chlorine and hydrochloric acid, since these produce a yellow color. 
(CI and FeCla are yellow.) 

A glass bulb or a small funnel is placed in each test-tube and the water in 
the bath then heated to boiling and boiled until all the carbonaceous matter 
has dissolved, the tubes being shaken from time to time to prevent formation 
of a film of oxide. Low-carbon steels require about twenty minutes, whereas 
steels of over 1% carbon require about forty-five minutes. (Blair.) As soon 
as the bubbles cease and the brownish flocculent matter disappears, the rack 
is pranoved from the bath and placed in a casserole of cold water. (Prolonged 
heating and strong light each causes fading of the color due to combined carbon.) 



^ 



t r 
Fio. 25. — Carbon Tubes. 




Fig. 26. — Color Comparator or Camera. 



Color Comparison. This is made in graduated, clear, colorless, glass cyl- 
inders called carbon tubes. The form shown in Fig. 25 was found by the 
writer » to be the most satisfactory type for a steel-works laboratory where rapidity 
of manipulation was essential. The bend at the upper portion of the tube 
facilitates mixing of the solution upon dilution with water, the tube being tilted 
back and forth until the solution is homogeneous, the bend preventing the 
liquid from spilling. The dilution should be at least twice that of the amount 
of nitric acid used, as this amount of water is necessary to destroy the color 
due to ferric nitrate. 

The standard is poured into the carbon tube and the rinsings from the test- 
tube added. The solution is diluted to a convenient multiple in cc. of the 
carbon content. For example, 0.45% carbon sample may be diluted to 9 cc, 
then each cc. will represent 0.05% carbon. The sample is placed in a second 
tube of exactly the same diameter, wall thickness, and form. If the solution 
of the sample is darker than the standard, water is added little by little, 



* W. W. Scott. 



110 CARBON 

followed by mixing, until the shade matches the standard. If the standard, 
on the other hand, is darker than the sample, a greater dilution of the standard 
is necessary, the cc. again representing a multiple of the carbon content. For 
example dilution of the .45% carbon sample to 15 cc. makes each cc. to repre- 
sent 0.03 carbon. (It is frequently advisable to take a standard of lower carbon 
content in place of greater dilution of the standard.) 

Example. Suppose in the first case the dilution of the sample was 15 cc. 
in order to match the standard, then 15X0.05=0.75% carbon. Six cc. dilution 
case 2 =6X0.03 =0.18% carbon. 

The color comparison can be best made in a "camera," a long box with one 
end closed by a ground-glass screen, Fig. 26. Parallel to the screen and near 
it, two holes through the top of the box admit the test-tubes. The inner walls 
of the camera are blackened to prevent reflection of light. If a camera is not 
available, the tubes may be held side by side and compared against a sheet 
of white paper held as a background. 

ANALYSIS OF GRAPHITE 
Determination of Carbon 

The procedure for determining carbon in graphite is the same as that de- 
scribed for determination of carbon in difficultly combustible organic substances. 

The material is broken down in a steel mortar and powdered in an agate 
mortar. About 0.2 gram is taken for the determination and mixed with copper 
oxide to assist the combustion, then placed in the boat and the combustion 
of the carbon carried on according to the standard method in the combustion 
tube. 

CO,X0.2727=C. 

VOLUMETRIC DETERMINATION OF HYDROCYANIC ACID^ 

The method depends upon the decolorization of the blue anunoniacal solu- 
tions of cupric salts by a soluble cyanide, the reduction to cuprous condition 
being available for an accurate quantitative estimation of the cyanide. 

Standard Copper Sulphate. Twenty-flve grams of copper sulphate, 
CUSO4 -51110 are dissolved in a 1000-cc. flask with 500 cc. of distilled water and 
ammonium hydroxide added until the precipitate that first forms dissolves and 
a deep blue solution is obtained. Water is now added to make the volume exactly 
1000 cc. The cupric solution is standardized by running a portion into a solu- 
tion containing 0.5 gram pure potassium cyanide, KCN, per 100 cc. of water 
and 5 cc. of ammonium hydroxide until a faint blue color is evident. Chlorides 
do not interfere. 

Procedure. 0.5 gram of the soluble cyanide is dissolved in 100 cc. of water 
and 5 cc. strong anunonium hydroxide added. The standard cupric sulphate 
solution is now added until the blue color is obtained. The cc. required mul- 
tipUed by the factor of the copper salt in terms of the salt sought gives the 
weight of that salt in the sample. 

Note. Test for Cyanide. This depends upon the solvent action of HCN upon 
freshly precipitated HgO in presence of KOH. The filtrate is tested for mercury 
in an acid solution by addition of HsS. (Hood.) 

» J. McDoweU, C. N., 1904, p. 221. 



CARBON 111 



Uebig's Method for Determination of HydroQranic Acid. Soluble 

Cyanides ^ 

Silver nitrate reacts with an alkali cyanide in neutral or alkaline solution as 
follows: AgN0,+2KCN = Ag(CN) JC+KNO,. The potassium silver cyanide is 
soluble, hence the precipitate that first forms immediately dissolves on stirring 
as long as the cyanide is present in excess or in sufficient quantity to react 
according to the equation. A drop of the silver salt in excess will produce 
a permanent turbidity, owing to the following reaction: 

Ag(CN)JC+AgN0,=2AgCN+KN0„ the insoluble AgCN being formed. 

Procedure. The alkali cyanide contained in a beaker placed over a sheet 
of black glazed paper, is treated with 4 to 5 cc. of 10% KOH solution and 
diluted to 100 cc. The liquid is now titrated with standard silver nitrate, with 
constant stirring, imtil a faint permanent turbidity is obtained. 

One cc. N/10 AgNO,= 0.0013022 gram KCN. 

For his review and criticism of this chapter the author wishes to mention Mr. J. M. 
Cratty, Chief Chemist, U. S. Navy Yards, Philadelphia, Pa. 

^ Ann. d. Chem. und Pharm., 77, p. 102. 



CERIUM AND THE OTHER RARE EARTHS 



R. Stuart Owens * 



Group- 



Symbol. 



YUrium Yt 

Erbium Er 

Hclmium Ho 

Thvlium Tm 

Dysprosium D y 

YUerbium Yb 

(Neo-ytterbium) 

Lutecittm Lu 

Europium Eu 

Vidorium 

Group 2: 

Terbium Tb 

Gadolinium Gd 

Group 3: 

Cerium Ce 

Lanihanum La 

Neodymium Nd 

Praseodymium Pr 

Samarium Sa 

Scandium So 

Decipium 



At. Wt.t 



88.7 



167 
163 
168 
162 
173 



7 
5 
5 
5 
5 



175.0 

152.0 

Discovcrv 

159.2 
157.3 

140.25 
139.0 
144.3 
140.9 
150.4 
44.1 
Discovery 



Sp. Gr. 



3.800 
4.770 



not confirm 



1.310 

6.625 
6.163 
6.544 
6.544 
7.700 

not confirm 



M. P. 



1250 



1800 



ed. 



950 



ed. 



840 

940 

1350 

1300 



Oxides. 



Yt,0, 
Er,0,, ErA 

Tm,0, 

Yb,0, 



TbjO, 
Gd,0, 

Ce,0,, CejOt 

Nd,0, 
Pr,0, 
Sm,0, 
SciOi 



• Accordiig to Bohm (Browning, *' Introduction to the Rarer Elcmenta.") 
t International atomic weights, 1016. 

DETECTION 

The samples having been brought into solution by one of the methods 
detailed under preparation and solution of the sample, the elements may be 
detected by one of the following tests: 

Spectroscopic. Many of the rare earth's elements have either character- 
istic spark spectrums or absorption spectrums and their presence may be detected 
by this means. 

Yttrium, no absorption spectrum; gives brilliant spark spectrum. 

Erbium, 

Ytterbium. 

Terbium, 

Cerium, 

Lanthanum, 

Samarium, 

Scandium, 

Praseod>Tnium, no 

Neo<l>TTiium, no 

Cerium shows lines of greatest intensity in the arc spectrum at 4337.9, 
4527.5, 4386.9, 4594.1. In the spark spectrum at 4400.3, 4562.5, 4572.4, 4594.1, 
4628.3. All of these lines are in the visible spectrum. 

* Research Chemist, New York City. 

112 



gives 
no 




no 




no 




no 




gives 
no 




no 




»% ^^ 







i I 


H H 11 




< < 


It H ft 




i i 


I i 1 1 (t 




< < 


a tt tt 




( i 


ii (( tt 




no 


Spark spectrum. 




i i 


11 n 




1 1 


n tt 



CERIUM AND THE OTHER RARE EARTHS 113 

In the wet way cerium may be detected when in the form of cerium nitrate 
by boiling with lead peroxide and nitric acid. A deep yellow color is imparted 
to the solution, due to the formation of eerie nitrate. 

Cerium may be detected by the addition of sodium hypochlorite to the solu- 
tion of a colorless cerous salt. Red eerie hydroxide is precipitated. The test 
may be confirmed by the chlorine gas evolved when the precipitate is dissolved in 
hydrochloric acid. 

Cerous salts are precipitated by fixed alkalies and are insoluble in excess. 
Tartaric acid hinders the precipitation. Ammonium sulphide also precipitates 
the hydroxide. Oxalic acid precipitates cerous oxalate, white, from moderately 
acid solutions. It is soluble in hot ammonium oxalate but precipitated by 
dilution with cold water. 

Lanthanum may be detected by adding iodine to the washed precipitate, 
formed by the addition of ammonium hydroxide to a solution of its salts. A 
characteristic blue coloration results. 

Praseodymium, neodymium, may be detected by the reddish color of their 
solutions also by the rose red or violet color imparted to a bead of microcosmic 
salt when heated in the flame of a blow pipe. 

Scandium. The hydrochloric acid solution of a scandium salt, when boiled 
for thirty minutes with solid NazSiFU j?ives a precipitate which is free from all 
the other rare earths, the scandium taking the place of the sodium in the com- 
pound. 

Ytterbium may be detected by adding to a neutral solution H2SeOs*4H20. 
A white precipitate of Yb2(Se03)s, which is insoluble, results. 

Erbium. In the flame test this earth gives an intense green light. 

ESTIMATION 

The estimation of the rare earths is not required, other than Cerium, at 
the present time except in a few special instances as the various elements have 
found but limited commercial applications. They have all been separated 
from their native combinations, but only a few have been isolated and many are 
still believed to be combinations of elements. 

Cerium enters into the manufacture of Welsbach mantles; in the form of 
Ce2(S04)s it is used in the manufacture of aniline black; as oxalate, it is used in 
medicine, and as metal in alloys. 

Yttrium is employed in the fabrication of Nemst lamp filaments and gas 
mantles. 

The most important ores which contain the rare earth elements are: 

Monazite, (Ce, La, Di, Th)xP04, raw material for Ce, La. 

Gadolinite, BciFeYiSiaOio, 

Xenotime, YtP04, 

Euxenite, R'"(NbO,), • R/''(TiO,)s •3/2H,0, 

Cerite, (Ca, Fe)(CeO)(Ce,.30H)(SiO,),, 

Samarskite, R/'R2"'(Nb, Ta)«02i, 

Yttrotantalite, R"R,'"(Nb, Ta)40i6.4H20, 

Sipylite, complex, 

Keilhauite, complex silicate. 

In the formulas given above R" stands for any dibasic radical or element 
while R'" stands for any tribasic radical or element. 







Yt earths. 






Yt 








Yt 








Ce 








Yt 








Yt 








Yt 








Yt 





114 CERIUM AND THE OTHER RARE EARTHS 



Preparation and Solution of the Sample 

1. Fusion Method. The finely pulverized sample is fused with sodium 
carbonate and the melt after cooling is extracted with cold water. A sufficient 
quantity of hydrochloric acid to impart an acid reaction is added. The solu- 
tion obtained is evaporated to dryness and baked to dehydrate the silica, then 
treated with a little hydrochloric acid and after dilution with water, filtered. 
Ammonia water is added to the solution in slight excess and the solution allowed 
to stand until the precipitate has settled. It is then filtered off, washed with cold 
water and dissolved in hydrochloric acid. AD of the rare earths are then present 
in the solution as chlorides. 

2. Acid Ebctraction. Decomposition of the finely pulverized sample may 
be effected by mixing it with a sufficient quantity of sulphuric acid to make a 
paste and then heating the mass, slowly at first and then gradually increasing 
the heat to dull redness when fumes of S0» appear. After cooling, the mass 
is extracted with cold water and the metals of the HjS group removed in the usual 
way. The rare earths are then present in the solution as sulphates and may be 
separated by one of the methods detailed below. 

3. Acid Extraction. A strong mixture of nitric acid and hydrochloric acid 
may be used to effect the decomposition of some of the minerals. The solution 
after being evaporated to dryness and baked leaves a residue which contains 
the mixed rare earths. The rare earths are dissolved in dilute hydrochloric acid. 
Any silica present is filtered off and the rare earths present in the clear solution 
may be separated by one of the methods detailed below. 

4. Decomposition by Means of Hydrofluoric Acid.* Samarskite and 
euxenite in the finely powdered state are moistened with their own weight of water 
and twice as much fuming hydrofluoric acid. The attack takes place in a few 
seconds. When the violent action is over the solution is evaporated to drjrness 
on the steam bath, taken up with water (30 to 40 cc. for a 5-gram sample) and 
the contents of the dish filtered and washed. The mineral is then divided into 
two portions, the filtrate containing all the metallic acids, iron and manganese, 
the insoluble portion containing all the rare earths and uranic acid. 

The difficulty of attack increases in proportion to the amount of tantalic 
acid present in the sample. The rare earths are then extracted from the incoluble 
portion by one of the methods previously mentioned. Fusion with sodium car- 
bonate is preferred. 

SEPARATIONS 

Separation of the rare earths from iron, aluminum and thorium * may 

be effected by adding sodium fluoride to the hydrochloric acid solution of the 
Iron Group ^whlch has been precipitated as hydroxide. The precipitate, which 
consists of the double fluorides of the rare earths and thorium, is washed thoroughly 
and evaporated with sulphuric acid on the sand bath to decompose the fluorides. 
This process removes the alkaline earths as insoluble sulphates. The excess acid 
is removed by fuming and the solution of the sulphates after diluting and warm- 
ing is treated with sodium thiosulphate in solution. Thorium thiosulphate is 
precipitated. In solution are the sulphates of all the rare earths. Scandium 

^ Prescott^d Johnson. 

s Browaip^s "Introduction to the Rarer Elements." 



.•• ) 



CERIUM AND THE OTHER RARE EARTHS 115 

is also precipitated as thiosulphate if the solution in sulphuric is fumed too long 
and a neutral solution results. 

Calcium and manganese, which may also come down with an oxalate pre- 
cipitate of the rare earths, may be separated from the earths of the yttrium group 
by precipitation of the solution with oxalic acid, filtering off the precipitate, dis- 
solving it in nitric acid and evaporating to dr3mess to decompose the manganese 
salts. Extracting with water leaves the manganese in the residue. Treat the 
filtrate with ammonia water. The yttrium group precipitates as hydroxides 
and may be filtered from the calcium, which remains in solution. 

Cerium, lanthanum, praseodymium, neodymium, europium and gado- 
linium may be separated from the other rare earths by adding a saturated solu- 
tion of potassium sulphate to the sulphate or chloride solution of all of the 
rare earths. The above-mentioned elements form insoluble double salts. 

Scandimn may be separated from yttrium by boiling a solution of the nitrates. 
A basic scandium nitrate is first precipitated. 

Yttrium Group. Barium carbonate forms no precipitate in the cold, hence 
the elements comprising same may be separated from aluminum, iron, chromium, 
thorium, cerium, lanthanum, praseodymium, and neodymium by this means. 

Yttrium Group. The precipitation of the group as hydroxides is not affected 
by the presence of tartaric acid. Hence the members may be thus separated 
from aluminum, gluciniun, thorium, zirconium, and iron. 

Praseodymium, neodymium, lanthanum, and samarium may be separated 
from each other by the fractional precipitation of a dilute solution of the nitrates 
with a very dilute solution of ammonia water (1 gram of NH3 in 500 cc). The 
first precipitates are rich in samarium; the didymiums come down next and 
the lanthanum in the last portions. By a continual repetition nearly pure salts 
may be obtained. 

Besides the separations mentioned above the group members may be freed 
from each other by various other methods, as for example: 

(1) Fractional crystallization of the picrates. 

(2) Fractional crystallization of the double magnesium nitrates. 

(3) Fractional precipitation of the oxalates in a nitric acid solution, etc. 



GRAVIMETRIC ESTIMATIONS 

Owing to the fact that the quantitative separation of the rare earths is only 
accomplished by the expenditure of a vast amount of time and labor and that 
the various elements with the exception of but few have found no commercial 
application, exact methods have not been worked out for the various quanti- 
tative assays. 

Cerium, however, which is the most important, may be determined as 
follows: 

The element having been brought into solution by one of the methods detailed 
above and separated from the base metals, silica and thorium may be isolated 
from the other rare earths by precipitation in a slightly acid solution with oxalic 
acid. The precipitate is allowed to settle twenty-four hours, filtered, washed 
with water and ignited. The oxides are then dissolved in hydrochloric acid and 
precipitated as hydroxide by the addition of an excess of caustic potash. The 



116 CERIUM AND THE OTHER RARE EARTHS 

precipitate, suspended in solution, is subjected to the action of chlorine gas 
which is bubbled through in a steady stream. All of the rare earths except 
cerium are converted into the chlorides, while the latter remains as a reddish, 
gelatinous precipitate, eerie hydroxide — (Ce(0H)4). This may be filtered off, 
washed, ignited and weighed as oxide (Ce02). 

Cerium may be determined in its salts by precipitation with oxalic acid, 
allowing to settle out, filtering, washing, and igniting to the oxide. 

VOLUMETRIC METHOD FOR THE DETERMINATION 

OF CERIUM 

Method of Franz Stolba.^ The cerium having been separated from all 
of the other rare earths by some procedure, as, for example, that outlined above 
under Gravimetric Determination, may be precipitated as eerie oxalate. (Dis- 
solving the hydroxide in hydrochloric acid and then precipitating with oxalic 
acid.) The oxalate precipitate is filtered off, washed with water until free from 
hydrochloric acid, and transferred to a beaker containing a small quantity of 
sulphuric acid and a sufficient quantity of water. The mixture is warmed to 
about 70** C. and titrated with a standard solution of KMnOi. During the 
process of titration the quantity of undissolved matter diminishes and the 
change of color at the end is very distinct. The solution of KMnOi used is pre- 
viously standardized, using a known amount of pure eerie sulphate and the same 
quantities of water and sulphuric acid. 

Determination of Cerium in Welsbach Mantles. 

Colorimetric Method^ 

Bum off the organic matter and heat with about three times their own wei'?ht 
of H2SO4 (cone.) on a sand bath. Allow to cool and pour into 20 cc. of wat3r. 
In twenty-four hours the sulphates are completely dissolved and the solution 
after neutralizing the excess of acid with ammonia water is precipitated with 
oxalic acid. The oxalates after settling out are filtered, washed, transferred to 
a porcelain casserole and digested with nitric acid, a little being added at a time 
until complete decomposition has taken place. Evaporate to dryness to remove 
the excess acid. The nitrates of cerium and thorium are dissolved in water 
and made to volume. Aliquots are then taken and diluted in comparison tubes, 
1 cc. HjOi (Merck^s perhydrol) is added. On adding ammonia water Th(OH) 
is colored orange in proportion to the amount of cerium present. In dilute 
solutions citric acid prevents the precipitation of the hydroxides and the color 
can be easily compared with standards containing known amounts. 

* Crooke's "Select Methods of Analysis." 

« Method of E. Benz, Z. angew Chem., 16, 300, 1902. 



CERIUM AND THE OTHER RARE EARTHS 117 



RARE EARTH OXALATES 

Conrert into the sulphates by evaporating with sulphuric acid, 
solid sodium sulphate in excess to the nearly neutral solution. 



Dissolve in water and add 



(1) Precipitate: 

Th, Ce, La. Pr, Nd. Sa, £u, Gd, etc., as double sodium sul- 
phates. 

Boil with an excess of sodium hydroxide, filter, wash with hot 
water and dissolve in nitric acid. ^ Treat the nitrates with 
an excess of sine oxide and potassium permanganate. 



(3) Pbkcipitate: 

CeO« andThO«-MnOi is pres- 
ent is removed by solution in 
hydrochloric acid and then 
precipitation of the Ce and Th 
as double sulphates with so- 
dium. 

Precipitate is boiled with an 
excess of sodium hydroxide, 
washed with hot water and 
dissolved in nitric aoid. Add 
ammonia and ammonium ox- 
alate and ammonium acetate. 



Precipitate: 

Cerium oxalate. 
Treat with ex- 
cess of sine oxide 
and solution of 
potassium per- 
manganate. Ce- 
rium oxide pre- 
cipitates. Dis- 
solve in hydro- 
chloric acid and 
precipitate c«- 
rium us oxalate 
with oxalic acid. 



Filtrate: 

Thorium 
oxalate is 
treated 
with am- 
monia in 
excess, ig- 
nite. 
ThOi. 



(3) Filtrate: 

Saturate the solution with 
sodium sulphate and wash 
the precipitate forme<l 
with a solution of sodium 
sulphate. 



Precipitate: 

Boil with excess 
of sodium hy- 
droxide, filter 
and wash with 
hot water. 
Dissolve in a 
known amount 
of nitric acid 
and^ add an 
equivalent 
amount of 
magnesium ni- 
trate in solu- 
tion. Evapo- 
rate the solu- 
tion until upon 
blowing on 
surface small 
crystals form. 
Spray water on 
surface and al- 
low to crystal- 
lise. 

Lanthanum 
crystallises Ist, 
Prnteodymium 
crystallises 2d, 
Neodymium 
crystallises 3d 
Samarium 
crystallises 4th 
Europium 
crystallises 5th 
Gadolinium 
crystallises 6th 

The crystallisa- 
tion is con- 
trolled by the 
spectroscope. 
The greater 
number of 
times the 
earths arc frac- 
tionated the 
purer the prod- 
uct will be. 



Fil- 
trate: 

Combine 
with fil- 
trate 
No. 1. 



(I) Filtrate: 

As double sodium sulphates. Yt, Tb, 
Dy, Ho. Er, Tm, Yb, Sc, etc. 

Add oxalic acid in excess to precipitate 
the earths as oxalates. 



(2) Precipitate: 

As an excess of sulphuric 
acid and evaporate to form 
anhydrous sulphates of Yt, 
Tb, drous sulphates of Yt, 
Yb, Dy, Ho. Er, Tm. Sc, 
Yb. etc. 

Dissolve in cold water and 
pour over an excess of 
barium bromate. Stir 
well and place on the hot 
water bath . W hen double 
decomposition is complete 
(when the liquid gives no 
further precipitate with 
barium bromate solution 
after diluting and boiling), 
the mass is filtered and 
evaporated to crystallisa- 
tion. 

Terbium crystallises Ist, 
Dysprosium crystallises 
2d, Holmium crystallises 
3d, Yttrium crystallises 
4th, Erbium crystallises 
Thulium 5th, crystallises 
0th 

The mother liquor is made 
neutral with ammonia and 
naturated with potassium 
sulphate. 



Precipi- 


Filtrate: 


tate 

Scandium. 
Potassium 
Sulphate. 


Add oxalic 
acid. 

Ytterbium is 
precipitat- 
ed as oxal- 
ate. 



(2) Fil- 
trate: 

Add to 
filtrate 
No. 3. 



CHLORINE 

ft 

Wilfred W. Scott and Wm. F. Doerflingbr 

CluaUwU 35.46; D. (air), 2.491; m.p. -lOl.S'*;^ b.p. —SS.e^'C.; oxides, 

CI2O, CIO2, CliOj. 

DETECTION 

Free Chlorine. The yellow gas is recognized by its characteristic odor. 
It liberates iodine from iodides; it bleaches litmus, indigo, and many organic 
coloring substances. 

Chlorides. Silver Nitrate Test. In absence of bromides and iodides, 
which also form insoluble silver salts, silver nitrate precipitates from solutions 
containing chlorides white, curdy, silver chloride, AgCl (opalescent with traces), 
soluble in NH4OH (AgBr slowly soluble, Agl difficultly soluble), also soluble 
in concentrated ammonium carbonate (AgBr is very slightly soluble; Agl is insol- 
uble). Silver chloride turns dark upon exposure to light. 

Free Hydrochloric Acid. Manganese Dioxide, Potassium Permanganate, 
and certain oxidizing agents liberate free chlorine gas when added to solutions 
containing free hydrochloric acid. The gas passed into potassium iodide lib- 
erates free iodine, which produces a blue solution with starch. 

Concentrated Sulphuric Acid added to chlorides and heated liberates HCl 
gas, which produces a white fume in presence of ammonium hydroxide. 

Detection in Presence of Cyanate, Cyanide, Thiocyanate. An excess 
of silver nitrate is added to the solution, the precipitate filtered off and boiled 
with concentrated nitric acid to oxidize the cyanogen compounds and the white 
precipitate, silver chloride, subjected to the tests under chlorides to confirm 
the compound. 

Detection in Presence of Bromide and Iodide. About 10 cc. of the 
solution is neutralized in a casserole with acetic acid, adding about 1 to 2 cc. 
in excess, and then diluting to about 6 vohimes with water. About half a 
gram of potassium persulphate, KzSjOg, is added and the solution heated. 
Iodine is liberated and may be detected by shaking the solution with carbon 
disulphide, which is colored blue by this element. Iodine is expelled by boiling, 
the potassium persulphate being repeatedly added until the solution is colorless. 
Bromine is liberated by adding 2 or 3 cc. of dilute sulphuric acid and additional 
persulphate. A yellowish-red color is produced by this element. Carbon 
disulphide absorbs bromine, becoming colored yellowish red. Bromine is expelled 
with additional persulphate and by boiling. The volume of the solution should 
be kept to about 60 cc, distilled water being added to replace that which is 
expelled by boiling. When bromine is driven out of the solution, the silver 
nitrate test for chlorides is made. A white, curdy precipitate, soluble in 
ammonium hydroxide and reprecipitated upon acidifying with nitric acid, is 
produced, if chlorides are present. 

1 Ref. Cis. 35 (2d Ed.), U. S. Bureau of Standards. 

118 



CHLORINE 119 

If Chlorates are Present. The halogens are precipitated with silver nitrate, 
the precipitate dissolved with zinc and sulphuric acid and the solution treated 
as directed in the preceding paragraph. 

Test for Hypochlorite. Potassium hypochlorite, KCIO, shaken with mer- 
cury forms the yellowish-red compound HgiOCU,* which does not form with the 
other potassium salts of chlorine, i.e., KCl, KCIO», KClOi, KCIO4. 

Hypochlorites decolorize indigo, but do not decolorize potassium perman- 
ganate solutions. If arsenious acid is present, indigo is not decolorized until 
all df the arsenious acid has been oxidized to the arsenic form. 

Tests for Chlorite. Potassium permanganate solution is decolorized by 
chlorites. (The solution should be dilute.) 

A solution of indigo is decolorized, even in presence of arsenious acid (dis- 
tinction from hypochlorites) . 

Detection of Chlorate. The dry salt heated with concentrated sulphuric 
acid detonates and evolves yellow fumes. 

Chlorates liberate chlorine from hydrochloric acid. 

Perchlorate. The solution is boiled with hydrochloric acid to decompose 
hypochlorites, chlorites and chlorates. Chlorides are removed by precipitation 
with silver nitrate, the filtrate evaporated to dryness, the residue fused with 
sodium carbonate to decompose the perchlorate to form the chloride, which may 
now be tested as usual. 

ESTIMATION 

The determination of chlorine is required in a large number of substances. 
It occurs combined as a chloride mainly with sodium, potassiiun and mag- 
nesium. Rock salt, NaCl, sylvine, KCl, camallite, KCl • MgCU • 6H2O, matlockite, 
PbCU-PbO; horn silver, AgCl, ataeamite, CuCl2»3Cu(OH)2, are forms in which it 
is found in nature. Chlorine is determined in the evaluation of bleaching powder. 
It is estimated in the analysis of water. 

Preparation and Solution of the Sample 

In dissolving the sample the following facts should be borne in mind: 
Although chlorides are nearly all soluble in water, sUver chloride is practically 
insoluble (100 cc. dissolves 0.000152 gram at 20° C); mercurous chloride is 
nearly as insoluble as silver chloride (0.00031 gram); lead chloride requires heat 
to bring it into solution (in cold water only 0.673 gram soluble per 100 cc. of 
water). Chlorides of antimony, tin, and bismuth require free acid to keep 
them in solution. Hydrochloric acid increases the solubility of silver, mercury, 
lead, antimony, bismuth, copper (Cu'), gold and platinima, but decreases the 
solubility of cadmium, copper (Cu")> nickel, cobalt, manganese, barium, cal- 
cium, strontium, magnesium, thorium, sodiiun, potassium and anmionium chlorides. 

Chlorine gas is most readily dissolved in water at 10° C. (1 vol. HjO dissolves 
3.095 vols. CI). Boiling completely removes chlorine from water. 

Hypochlorites, chlorites, chlorates, and perchlorates are soluble in water. 

The chlorine may be present either combined or free. In the combined 
state it may be present as free hydrochloric acid or as a water-soluble or insol- 
uble salt. 

^ Presoott and Johnson, Qual. Chem. Anal. D. Van Nostrand Co. 



120 CHLORINE 

Water-soluble Chlorides. Chlorides of the alkali or alkaline earth groups 
may be treated directly with silver nitrate upon making slightly acid with 
nitric acid, the chlorine being determined either gravimetrically or volumetrically 
according to one of the procedures given later. It is convenient to work with 
samples containing 0.01 gram to 1 gram of CI. The sample is dissolved in about 
150 cc. of water, made acid with nitric acid with about 5 to 10 cc. in excess of 
the point of neutralization, should the sample be alkaline. Then the chlorine 
combined as chloride is determined as directed later. 

If the water solution contains a chloride of a heavy metal which forms 
basic salts (e.g., stannic, ferric, etc., solutions), or which may tend to reduce 
the silver solution, it is necessary to remove these by precipitation with ammo- 
nium hydroxide, or by sodiiun hydroxide, or potassium carbonate solution. 
The salt is dissolved in water and acidified with HNOi, adding about 10 cc. in 
excess, for about 150 cc. of solution. (This excess HNO» should be suflficient 
to oxidize substances which would tend to reduce the silver reagent; e.g., FeS04, 
etc.) Ammonia solution (free from chloride) is added in sufficient quantity 
to precipitate the heavy metals iron, manganese, aluminum, etc. The mixture is 
filtered and the residue washed several times with distilled water. Chlorine 
is determined in the filtrate by acidifying with HXOs as directed above. 

Water-insoluble Chlorides. The chloride may frequently be decomposed 
by boiling with sodium carbonate solution. Many of the minerals, however,, 
require fusion with sodium carbonate to prepare them for solution; e.g., apatite, 
sodalite, etc. Silver chloride may also be decomposed by fusion. 

Silver Chloride. The sample is mixed with about three times its weight of 
NaiCOi and fused in a porcelain crucible until the mass has sintered together. 
The soluble chloride, NaCl, is leached out with water, leaving the water-in- 
soluble carbonate of silver, which may be filtered off. The filtrate is acidified 
with HNOi and chlorine determined as usual. 

Chlorine in Rocks. The finely ground material is fused with about five times 
its weight of p>otassium carbonate. The melt is extracted with hot water, cooled 
and the solution acidified with nitric acid (methyl orange indicator), and the 
solution allowed to stand several hours (preferably over night). If silicic acid 
precipitates, the solution is treated with ammonia and boiled, filtered and the 
filter washed with hot water. The cooled filtrate is acidified with nitric acid and 
chlorine determined as usual. If silicic acid does not separate, the addition of 
anunonia may be omitted and chlorine determined in the solution. 

Free Chlorine. Free chlorine may be determined volumetrically according 
to the procedure given under this section. If it is desired to determine this 
gravimetrically, a definite amount of the chlorine water is transferred by means 
of a pipette to a flask containing ammonia solution and the mixture heated 
to boiling. The cooled solution is acidified with nitric acid and the chloride 
precipitated with silver nitrate according to the standard procedure given 
on page 127. 

Note. Free chlorine cannot be precipitated directly, as the following reaction 
takes place: 6Cl-|-6AgNO,-|-3HaO=5AgCl-|-AgCl()a-|-6H>fO,. 

Reaction of chlorine with ammonia: 2Cl-h2NH40H = NH4Cl-hNH40Cl-hH,0. 
WTien the solution is boUed, NH4OCI breaks down, e.g, 3NH40Cl-f2NH,=3NH4Cl 
-fNj-hSHjO. 

Chlorine in Ores and Cinders. One hundred grams of the finely ground 
ore or cinder are placed in a 500-cc. flask, containing 300 cc. of strong sulphuric 



CHLORINE 121 

acid (Cl-free). The 'flask is shaken to mix the sample with the acid and then 
connected with an absorption apparatus, containing distilled water or dilute 
caustic solution. The sample is gradually heated, the distillation flask resting 
upon a sand bath. After two hours, which is sufl^cient to expel all the chlorine 
as hydrochloric acid, the contents of the absorption tubes are filtered, if free 
sulphur is present (sulphide ores), nitric acid added and the filtrate brought 
to boiling to oxidize any SOi that may be present. Chlorine is precipitated 
according to the standard procedure on page 124. 

During the run the distilling flask should be shaken occasionally to prevent 
caking. Suction applied at the absorption end of the apparatus and a current 
of air swept through the system aids in carrying over the HCl into the water 
or NaOH. 

Determination of Halogens in Organic Compounds. Method of 

Carius ^ 

Organic compounds may be decomposed by heating with strong nitric acid 
at high temperatures under pressure. If this heating is conducted in the presence 
of silver nitrate, the halogen hydride, formed by the action of nitric acid on the 
organic compound, is converted to the silver halide. This is weighed, or the 
excess AgNOa titrated (p. 125). Arsenic, phosphorus, and sulphur arc oxidized 
to arsenic, phosphoric, and sulphuric acids, the metals present being converted 
to nitrates. 

Procedure. About 0.5 to 1 gram of powdered silver nitrate Is introduced, 
by means of a glazed paper funnel, into a heavy- walled, bomlvglass tube, which 
is sealed at one end and is 50 cm. long, 2 cm. in diameter and about 2 mm. thick- 
ness of wall. About 30 cc. of strong nitric acid (96%), free from chlorine, are 
introduced by means of a long-stemmed funnel, to avoid wetting the upper portion 
of the tubing. About 0.1 gram of the organic substance, contained in a small bore, 
thin wall, glass tube closed at one end (4-5 cm. long), is introduced into the bomb 
tube, inclined to one side. The small tube should float in the nitric acid, as it is 
important that the material should not come in contact with nitric acid until the 
bomb has been sealed, as loss of halogen is apt to occur with open tubes. The 
upper end of the bomb is softened in the blast-lamp flame, drawn out to a thick- 
walled capillary tube and fused. 

When cold, the bomb is wrapped in asbestos paper, shoved into an iron tube 
of a bomb furnace and the heat turned on. The heating is so regulated that the 
temperature is raised to 200 ° C. in three hours. If a higher temperature is neces- 
sary, the heating should be such as to cause a rise of 50° C. in three hours. Sub- 
stances of the aromatic series require eight to ten hours heating at 250 to 300° C, 
while aliphatic substances may be decomposed at 200° C. in about four hours.* 
Occasionally it is necessary to relieve the pressure in a tube after heating to 200° C, 
before taking to a higher temperature, by softening the tip of the cooled bomb in 
a flame, allowing the accumulated gas to blow out, resealing and again heating 
to the desired temperature. Evidence of crystals or drops of oil in the glass tube 
indicate incomplete decomposition. When the bomb is cooled, it is removed by 

^ Ann. d. Chem. u. Pharm. (1865), 186, p. 129. 

* Treadwell and Hall, Anal, Chem., J. W iley & Son. P. C. R. Kingscott and R. S. G. 
Knight, Methods of Quant. Org. Anal. Longmans, Green & Co. (1914), Clowes and 
Coleman, Quant. Chem. Anal., 445, P. Blakiston's Son & Co., 1900. 



122 CHLORINE 

taking out the iron sheath from the furnace and inclining It so that the glass capil- 
lary tip slides partly out of the tube. (The eyes should be protected by goggles.) 
The point of the capillary is held in the flame until the tip softens and the gas 
pressure is released by blowing through a passage in the softened glass. When the 
gas has escaped, a scratch with a file is made below the capillary and the tip 
broken off by touching the scratch with a hot glass rod. The contents of the bomb 
are poured out into a beaker, the tube washed out with water and the combined 
solution made to about 300 cc. This is heated to boiling and then allowed to cool. 
The halide precipitate is filtered through a Gooch crucible, then dried and weighed, 
or by titrating the excess AgNOj by Volhard's method, the halide may be esti- 
mated. 

If pieces of glass should be present, the precipitates, AgCl or AgBr, are 
dissolved, in anunoniura hydroxide, filtered and reprecipitated by acidifying 
with nitric acid. Agl may be dissolved by means of dilute sulphuric acid and 
zinc. The excess zinc is removed, the glass washed free of iodine, dried and 
weighed and its weight subtracted from the original impure Agl, giving the weight 
of the pure silver iodide. 

Lime Method for Determination of Halogens in Organic Matter 

A layer of lime (free from chloride), about 6 cm. long, is introduced into 
a difficultly fusible gla.ss tube, closed at one end (35 cm. long and with 1 cm. 
bore), followed by 0.5 gram of the substance, and 6 cm. more of the lime. The sub- 
stance is thoroughly mixed by means of a copper wire with a spiral end. The 
tube is nearly filled with lime, and in a horizontal position, gently tapped to cause 
the lime to settle and form a channel above the layer. The tube is placed in a 
small carbon combustion furnace. The heat is turned on, so that the front end 
of the tube is heated to dull redness and then the end containing the substance. 
When the organic matter has been decomposed, the tube is cooled and the contents 
transferred to a beaker and the lime dissolved in dilute nitric acid (Cl-free). The 
carbon is filtered off and the halogen determined as usual in the filtrate. 

Should a sulphate be present in the mixture, organic matter will reduce it to a 
sulphide, so that AgS will be precipitated along with the halides. To prevent 
this, hydrogen peroxide is added to the solution which should be slightly alkaline. 
The mixture is boiled to remove the excess of H2O2 and is then acidified with 
nitric acid, the solution filtered and the halide determined in the filtrate. 

With substances rich in nitrogen, some soluble cyanide is apt to form. The 
silver precipitate containing the halides and the cyanide is heated to fusion. The 
residue is now treated with zinc and sulphuric acid, the metallic silver and the 
paracyanogen filtered off and the halides determined in the filtrate. 

Sodium Peroxide Method 

Organic compounds may l>e decomix)sed by sodium peroxide in an open 
crucible without recourse to a sealed tube, as Is required by the Carius method. 
The following is the procedure outlined by Pringsheim.* 

About 0.2 gram of substance in a small steel crucible is treated with a cal- 

> C. N., 1905, 91, 2372, 215. 



CHLORINE 123 

ciliated quantity of sodium peroxide.^ The crucible should be only two-thirds 
of its height fuU; this is put in a large porcelain crucible, in which a little cold 
water is carefuUy placed, so that the steel crucible stands out 1 to 2 cm. This 
latter crucible is covered with its own cover, in which is a hole through which 
an iron wire heated to redness can be introduced with the object of starting 
the combustion. As soon as the combustion is completed the whole is plunged 
into the water in the larger crucible. The porcelain crucible is covered with a 
watch-glass and heated gently imtil the whole mass is dissolved. This point is 
recognized when no more bubbles are given off and when there are no more 
particles of carbon which have escaped combustion. The steel crucible is then 
removed and washed' carefully; the solution is filtered and treated with an 
excess of sulphurous acid (to neutralize the alkaline liquid, and to reduce the 
oxidized products: bromic, iodic acids, etc.). The solution is acidulated with 
nitric acid, then made to a volume of about 500 cc, and the halogens precip- 
itated with silver nitrate and the precipitate washed, dried and weighed as 
usual. 

SEPARATIONS ^ 

Separation of Chlorine and the Halides from the Heavy Metals. Halides 
of the heavy metals are transposed by boiling their solutions with sodium car- 
bonate, the heavy metals being precipitated as carbonates and the halides going 
into solution as sodium salts. 

Separation of Halides from Silver and from Silver Cyanide. The silver 
salt is treated with an excess of zinc and sulphuric acid, the metallic silver 
and the paracyanogen filtered off, and the halides determined in the filtrate. 

Separation of the Halides from One Another. Separation of Chlorine 
from Iodine. The method depends up>on the fact that nitrous acid sets iodine 
free from dilute solutions containing a mixture of halogen salts, bromides and 
chlorides being unaffected. 

The solution of the chloride and the iodide in an Erlenmeyer flask is diluted 
to 400 cc. and 10 cc. of dilute sulphuric acid, 1 : 1, are added. The gas from 2 
grams of sodium nitrite is passed into the solution at the rate of about five bub- 
bles per second.* (Pure sodium or potassium nitrite may be added directly to the 
solution in the flask.) The liberated iodine is now completely expelled by boiling 
until the evolving steam no longer reacts upon litmus paper. Should a deter- 
mination of iodine be desired the evolved gas is absorbed in a hydrogen peroxide 
sodium hydroxide solution according to the procedure described under iodine. 



* Charge of sodium peroxide is judged as follows: 




Per cent C and O in material. 


Amount of sugar to add. 


Amount of NasOs required. 


Over 75 
30 to 75 
25 to 50 
Below 25 





i the wt. of sub. 
An equal weight 


18 times wt. of sub. 
16 times wt. of sub. 
16 times wt. of sub. 
16 times wt. of sub. 



• Attention is called to " Methods in Chemical Analysis," by F. A. Gooch for useful 
information on the separation of the halogens. 

' Nitrous acid is generated by addition of dilute H9SO4 to NaNOs, the add being 
added drop by drop through a thistle tube with glass stop-cock. 



124 CHLORINE 

The contents of the flask are treated with sUver nitrate and the precipitated 
silver chloride determined as usual. 

Separation of Chlorine and Bromine from Iodine. The procedure is 
similar to the separation of chlorine from iodine with the exception that a more 
dilute solution is necessary to prevent the volatilization of bromine with the 

iodine. 

The neutral solution containing the halogens is diluted to about 700 cc. and 
about 2 to 3 cc. of dilute sulphuric acid, 1 : 1, are added and a sufficient amount 
of pure sodium nitrite introduced or nitrous acid gas passed into the solution 
as directed above. The solution is boiled until colorless and imtil the evolved 
steam no longer acts upon litmus paper. About twenty minutes' boiling after 
the color of iodine has disappeared from the flask will completely eliminate 
iodine; in this case, however, water should be added to the flask to replace that 
evaporated before the solution has been reduced to a volume of less than 600 cc. 

For determination of bromine in the residue remaining in the flask, see the 
chapter on this subject, page 79. 

GRAVIMETRIC METHOD 

Determination of Chlorine Combined as Chloride by Precipitation 

as Silver Chloride 

The procedure is the reciprocal to the one for determination of silver; in this 
case the soluble silver salt is added to the sodium chloride solution, in which 
chlorine is to be determined. 

Procedure. To the nitric acid solution of the chloride, prepared according to 
directions given under ^'Preparation and Solution of the Sample," is added silver 
nitrate solution in slight excess, stirring during the addition of the reagent. (Stir- 
ring aids the coagulation of the AgCl and hastens settling. It is advisable to allow 
the precipitate to settle sufficiently to clear the upper portion of the solution 
in order to detect whether further precipitation takes place upon addition of 
more of the reagent.) The mixture is now heated until it is hot to the touch 
and then allowed to settle for half an hour or more, preferably in the dark. It 
is filtered through a weighed Gooch crucible, washing the precipitate by decan- 
tation several times with cold water, slightly acid with nitric acid, and then 
the precipitate transferred to the Gooch is washed free of silver nitrate (HCl 
test) with cold distilled water. The Gooch is dried for fifteen to twenty minutes 
at 100° C. and then at about 130° C. to constant weight. The sample is now 

weighed as AgCl. * 

AgClX 0.2474= CI. 

Notes. Free chlorine is converted to chloride according to the procedure given 
for preparinft the sample, and then determined according to the procedure given 
above. If chlorine and chlorides are both present in the solution and each is desired, 
the free chlorine is determined according to the volumetric procedure given later, 
and the total chloride detei mined giavimetrically, then free chlorine subtracted 
from total chlorine and the result taken as combined chlorine of the solution. 

* The silver chloride should be completely soluble in ammonia. If it is not, the 
product is impure. To separate it from SiOj, AljOj, and other impurities, the pre- 
cipitate is dissolved in ammonia, the solution filtered free from the impurities, 
and the AgCl reprecipitated by acidifying with nitric acid and adding a few drops 
of silver nitrate. 



CHLORINE 125 

If a paper filter is used in place of the Gooch crucible, the greater part of the 
precipitate is removed, the paper ignited separately, the reduced silver oxidized with 
HNOs, a drop or so HCl added, then evaporated oft, and the residue combined with 
major portion of AgCl and ignited gently until the salt begins to melt. 

VOLUMETRIC METHODS 

Determination of Chlorine in Acid Solution, Silver Thiocyanate 

Ferric Alum Method 

The method, devised by Volhard,^ is applicable to titration of chlorine in 
acid solutions, a condition frequently occurring in analysis, where the Silver- 
Chromate Method of Mohr cannot be used. The method is based on the fact 
that when solutions of silver and an alkali thiocyanate are mixed in presence 
of % ferric salt, the thiocyanate has a selective action towards silver, combining 
with this to form thiocyanate of silver, any excess of that required by the silver 
reacting with the ferric salt to form the reddish-brown ferric thiocyanate, which 
color serves as an indication of the completion of the reaction. An excess of 
silver nitrate is added to the nitric acid solution containing the chloride, AgCl 
filtered ofif , and the excess of silver titrated with the thiocyanate in presence of 
the ferric salt. 

Copper (up to 70%), arsenic, antimony, cadmium, bismuth, lead, iron, zinc, 
manganese, cobalt, and nickel, do not interfere, unless the proportion of the latter 
metals is such as to interfere by intensity of the color of their ions. 

Preparation of Special Reagents. N/10 Ammonium or Potassium Thia- 
cyanate Solution, About 8 grams of ammonium or 10 grams of potassium salt 
are dissolved in water and diluted to one liter. The solution is adjusted by 
titration against the N/10 silver nitrate solution. It is advisable to have 1 cc. 
of the thiocyanate equivalent to 1 cc. of the silver nitrate solution. Owing to 
the deliquescence of the thiocyanates the exact amount for an N/10 solution 
cannot be weighed. 

N/10 Silver Nitrate, This solution contains 10.788 grams Ag or 16.989 grams 
AgNOj per liter. The silver nitrate salt, dried at 120° C, or pure metallic silver 
may be taken, the required weight of the Jatter being dissolved in nitric acid 
and made to volinne, or 17.1 grams of the salt dissolved in distilled water and 
made to 1000 cc. The solution is adjusted to exact decinormal strength by 
standardizing against an N/10 sodium chloride solution, containing 5.846 grams 
of pure NaCl per liter. 

Ferric Indicator, Saturated solution of ferric ammonium alum. Should 
this not be available, FeS04 may be oxidized with nitric acid, and the solution 
evaporated with an excess of H2SO4 to expel the nitrous fumes. A 10% solution 
18 desired. Five cc. of either of these reagents are taken for each titration. 

Pure Nitric Acid, This should be free from the lower oxides of nitrogen. 
Pure nitric acid is dUuted to contain about 50% HNO3, and boiled until per- 
fectly colorless. The reagent should be kept in the dark. Dilute nitric acid 
does not interfere with the method. 

Procedure. To the solution, containing 0.003 to 0.35 gram chlorine, in 
combination as a chloride, is added sufficient of the pure HNOj to make the solu- 
tion acid and about 5 cc. in excess. To the solution, diluted to about 150 cc, 
is added an excess of standard silver nitrate reagent. The precipitated AgCi 

' Liebig's Ann. d. Chem., 190, 1; Sutton, ** Volumetric Analysis,'' 10 £d. Z. Anorg. 
Chem., 6S, 330, 1909. 



126 CHLORINE 

is filtered off and washed free of silver nitrate. The filtrate and washings are 
combined and titrated with standard thiocyanate.^ 

The filtrate from the precipitated chloride U treated with 5 cc. of the ferric 
solution,' and the excess silver determined by addition of the thiocyanate 
until a permanent reddish-brown color is produced. Each addition of the 
reagent will produce a temporary reddish-brown color, which immediately fades 
as long as silver uncombined as thiocyanate remains. The trace of excess 
produces ferric cyanate, the reddish-brown color of this compound being best 
seen against a white background. From this titration the amount of silver 
nitrate used by the chloride is ascertained. 

One cc. N/10 AgNO, =0.00355 gram CI or 0.00585 gram NaCl. 

Volumetric Determination of Chlorine in a Neutral Solution, 

Silver Chromate Method 

The method, worked out by Fr. Mohr, is applicable for determination of 
chlorine in water or in neutral solutions containing small amounts of chlorine; 
the element should be present combined as a soluble chloride. Advantage is 
taken of the fact that silver combines with chlorine in presence of a chromate, 
AgsCrOi being decomposed as foUows: AgiCr04+2NaCl=2AgCl+NajCrO«. 
When all the chlorine has gone into combination as AgCl, an excess of KsCr04 
immediately forms the red AgjCrO*, which shows the reaction of AgNOi with 
the chl ride to be complete. 

Reagents. Tenth Normal Silver Nitrate Solution. Theoretically 16.d89 
grams AgNOs per liter are required. In practice 17.1 grams of the salt are dis- 
solved per 1000 cc. and the solution adjusted against an N/10 NaCl solution 
containing 5.846 grams NaCl per liter. 

Potassium Chromate. Saturated solution. 

Procedure. To the neutral solution (made so, if necessary, by addition of 
nitric acid or ammonium hydroxide), are added 2 or 3 drops of the potassium 
chromate solution. A glass cell ' (or a 50-cc. beaker) is filled to about 1 cm. 
in depth with water tinted to the same color as the solution being titrated. 
The cell is placed on a clear glass plate half covering the casserole containing the 
sample. The standard silver solution is now added to the chloride solution 
from a burette until a faint blood-red tinge is produced, the red change being 
easily detected by looking through the blank, colored cell. 

One cc. N/10 KiCrO* =0.003546 gram CI. 

Notes. Chlorides having an acid reaction (AlCU) are treated with an excess of 
neutral solution of sodium acetate and then titrated with silver nitrate. 

Elements whose ions form colored solution with chlorine are precipitated from 
the solution by sodium hydroxide or potassium carbonate, and the filtrate, faintly 
acidified with acetic acid, is titrated as usual. 

^Time is saved by filtering, throuf^h a dry filter paper, only a portion of the mixture 
made to a definite volume, and titrating an aliquot portion. The first 10-15 cc. of the 
filtrate are rejected. 

•Upon addition of the ferric solution no color should develop. If a reddish 
or yellowish color results, more nitric acid is required to destroy this. The amount 
of nitric acid does not affect results when within reasonable limits. 

' Depr^, Analyst, 6, 123; also, Systematic Handbook of " Volumetric Analysis," P. 
A. Sutton. 



CHLORINE 127 

Free hydrochloric acid is neutralized with ammonium hydroxide and titrated. 

It is advisable to titrate the sample under the same conditions as those observed 
during standardisation. The solution should be kept to small bulk and low tem- 
perature for accuracy on accoimt of the solubility of the silver chromate. 

Free chlorine should be converted to a chloride before titration. This may be 
acoomt)l]8hed, as stated under preparation of the sample^ by boiling with ammonium 
hydroxide. Free chlorine may be determined by sweepms the gas, by means of a 
current of air, into a solution containing potassium iodide, the liberated iodine titrated 
by N/10 thiosulphate, NasStOs, and the equivalent chlorine estimated. 

Volumetric Determination of Free Chlorine 

The determination depends upon the reaction C1+KI=KC1+I. The iodine 
liberated by the chlorine is titrated with Na«S«Oa and the equivalent CI cal- 
culated. 

Procedure. A measured amount of the chlorine water is added to a 
solution of potassium iodide in a glass-stoppered bottle by means of a pipette, 
the delivery tip of which is just above the surface of the iodide solution. The 
bottle is then closed and the contents vigorously shaken. The liberated iodine is 
titrated with tenth-normal sodium thiosulphate (2Na«Si03+Is=2XaI+Na2S406). 
When the yellow color of the iodine has become faint, a little starch solution 
is added and the titration completed to the fading out of the blue color. 

One cc. N/10 Na^SjOa =0.003546 gram CI. 

Determination of Hypochlorous Acid in the Presence of Chlorine 

The determination depends upon the reactions: 

2KI+H0C1=KC1+K0H+I, and 2KI+C1,=2KC1+I,. 

The alkali liberated by hypochlorous acid and the total iodine are determined 
and the calculations made for each of the constituents. 

Procedure. A measured volimie of N/10 HCl is added to a potassium iodide 
solution. To this the sample containing the hypochlorous acid and chlorine 
are added. The liberated iodine is titrated with N/10 NajSsOj. (The addition 
of starch is omitted.) The colorless solution is treated with methyl orange 
indicator and the excess of hydrochloric acid is titrated with N/10 NaOH. 
The potassium hydroxide, produced by the action of the hypochlorous acid 
upon the iodide, requires half as much acid for neutralization as the volmne 
of thiosulphate reqxiired by the iodine set free by the hypochlorous acid. 

Calculatioii. The cc. back titration with NaOH are subtracted from the 
total cc. of HCl taken «cc. HCl required by NaOH liberated by HOCl=il. 
Then 2A cc. =cc. Na,S,0. required by the I liberated by HOCl. Cc. A X0.005247 
«gram HOCl. The total NajSjOa titration minus 2A cc. (due to the iodine 
liberated by HOCl) =cc. NajSjOj that are required by the iodine liberated by 
chlorine. The cc. thus required multiplied by 0.003546= grams chlorine in the 
sample taken* 

' Six parts AgCrO*, dissolve in 100,000 parts HaO at 15.5**.— W. G. Young, Analyst, 
18, 125. 



128 CHLORINE 

Gravimetric Determination of Chloric Acid, HCIO3, or Chlorates, 
by Reduction to Chloride and Precipitation as Silver Chloride 

Reduction of the Chlorate. Among the methods of reduction of chlorates 
the following deserve special mention: 1. Reduction with Sulphur (rus Acid,^ 
2. Ferrous sulphate. 3. Zinc. 

1. About 0.2 to 0.5 gram of the salt is dissolved in 100 cc. of distilled water 
and either SOs gas passed into the solution or sulphurous acid in solution added 
in excess. The solution is now boiled to expel SO2 and the chloride precipitated 
as AgCl in presence of free nitric acid. 

2. The sample in 100 cc. of distilled water is treated with 50 cc. of crys- 
tallized ferrous sulphate (10% solution), heated to boiling, with constant stirring, 
and then boiled for fifteen minutes. Nitric acid is added to the cooled solution, 
until the deposited basic ferric salt is dissolved. The chloride is now precip- 
itated as AgCl, as usual. 

3. The dilute chlorate solution is treated with acetic acid until it reacts 
distinctly acid. An excess of powdered zinc is now added and the solution 
boiled for an hour. Nitric acid is added to the cooled solution in sufficient 
quantity to dissolve the zinc remaining. The solution is filtered, if necessary, 
and the chloride precipitated as usual. 

Factors. AgCl X 0.855 = KCIO,, or X 0.2474 = CI. 

Note. In absence of cyanides, carbonates and acids decomposed and vola- 
tilized by hydrochloric acid, or oxides, hvdroxides and substances other than chlorates 
that may be decomposed or acted upon by this acid, evaporation of the salt with HCl 
and ignition of the residue, or addition of an excess of anmionium chloride,' and sub- 
sequent heating will give a residue of chloride, which may be determined as usual 
and the equivalent chlorate calculated. Method by L. Blangey. 

The methods may be used in determining chlorates in presence of perchlorates, 
only the former being reduced to chlorides. Outline of the procedure is given later. 

Oravimetric Determination of Perchloric Acid by Reduction to 

Chloride 

A perchlorate ignited with about four times its weight of ammonium chloride 
in a platinum dish may be decomposed to chloride. A second treatment is 
usually necessary to change the salt completely. Platinum appears to act as a 
catalyser, so must be added in solution if a porcelain crucible is used. 

Ftocedure. About 0.2 to 0.5 gram of potassium perchlorate is intimately 
mixed with about 2 grams of ammonium chloride in a platinum crucible, the 
latter then covered with a watch-glass and the charge ignited gently for one 
and a half to two hours, the temperature being below the fusing-point of the 
residual chloride (otherwise the platinum would be attacked). A second addi- 
tion of ammonium chloride is made and the mix again heated as before. The 
resulting chloride may now be determined as usual. 

Factors. AgCl X 0.9667 = KCIO4, X 0.2474 = CI. 



» Blattner and Brassucr, Chem. Zeit. Rep., 1900. 24, 793. 
* Perchlorates are decomposed by ignition with NH4CI in 



presence of platinum. 



CHLORINE 129 

Determination of Chlorates and Perchlorates in Presence of One 

Another 

( 1 ) A portion of the sample is treated with about twelve times its weight of 
ammonium chloride in a platinimi dish (or in a porcelain dish with the addition 
of 1 cc. of hydroplatinic acid), and the mixture heated according to the procedure 
given for perchloric acid (page 128). The resulting chloride is determined as 
usual. This is the total chlorine in the sample. 

(2) In a second portion the chlorate is reduced by means of SO2 or FeSO*, 
according to directions given for determination of chloric acid, and chlorine 
determined. The chlorine of this portion is subtracted from the total chlorine, 
the difference multiplied by 3.9075 =KC104. The chlorine of the second portion 
multiplied by 3.4563 =KC10,, or AgCl in (2) subtracted from AgCl of (1) 
and the difference multiplied by 0.9667 = KCIO4. AgCl of (2) multiplied by 0.855 = 
=KC10,. 

Determination of Hydrochloric, Chloric, and Perchloric Acids 

in the Presence of One Another 

(1) Total Chlorine. If the determination is made in the valuation of niter 
a 5-gram sample is fused with about three times its weight of alkali carbonate ^ 
or calcium hydroxide,' in a platinum dish, whereby all the chlorine compounds 
are converted to chlorides. If the compounds are present as alkali salts, fusion 
with anunonimn chloride in a platinum dish may be made and the total chlorides 
determined after dissolving the residue in nitric acid. 

(2) Chloride and Chlorate. If the estimation is being made in niter, 5 
grams of the salt are treated with 10 grams of zinc dust (Cl-free) in presence 
of 150 cc. of 1% acetic acid. The solution is boiled for half an hour, filtered, 
and the chloride determined. In a mixture of alkali salts of hydrochloric, chloric, 
and perchloric acids, reduction may be accomplished by passing in SO3 gas or 
by adding ferrous sulphate and boiling according to directions given for the 
determination of chlorate. The chloride now present in the residue is due to the 
reduced chlorate and to the original chloride of the sample. 

(3) The chloride of the sample is determined by acidifying the salt with 
nitric acid (cold) and precipitating as AgCl. 

Perchlorate. The chloride and chlorate in terms of chlorine are subtracted 
from total chlorine of (1) and multiplied by the factor for the salt desired. 

Chlorate. The chlorine of (3) is subtracted from chlorine of (2) and mul- 
tiplied by the factor for the compound desired. 

Chloride. The AgCl of (3^ is multiplied by the appropriate factor. 

Factors. AgCl X 0.2474= CI, or X0.2544=HC1, or X0.4078=NaCl, or 
X0.5202 =KC1. 

AgCl X 0.855 =KC10a, or X 0.9667 =KC104. 

ClX 3.4563 =KC10,, or X 3.9075 ^KCIO*, or X2.1027=KC1, or. X 3. 0028 = 
NaClO,, or X 3.4535 =NaC104, or X 1.6486 =NaCl. 

» Mennick, Chem. Zeit. Rep., 1898, 22, 117. 

* Blattner and Brasseur. Chem. Zeit. Rep., 1900, 24, 793. 



130 CHLORINE 

Determination of Chlorine, Bromine, and Iodine in the Presence 

of Each Other 

The procedure is Bekk's modification of Baubigny^s method.* 
Procedure. The halogens are precipitated with an excess of silver nitrate, 
filtered onto asbestos or glass wool, washed, dried, and weighed as total halogens 
as silver salts. A second portion is precipitated and the moist, washed silver 
salts (0.3 to 0.4 gram) are treated with a solution of 2 grams of potassium 
dichromate in 30 cc. of concentrated sulphuric acid at 95** C, and d^ested for 
thirty minutes. By this procedure the iodine is oxidized to hydriodic acid 
(HIO») and chlorine together with bromine is liberated in form of the free 
halogen. Toward the end of the reaction a stream of air is led through the 
solution to remove any chlorine and bromine. This is now diluted to 300 to 400 
cc, filtered, and the hydriodic acid reduced by adding, drop by drop, with con- 
stant stirring, a concentrated solution of sodium sulphite, Na^SOs, until a faint 
odor of SOs remains after standing ten minutes. (Under certain conditions an 
excess may result in a partial reduction of the silver iodide.) The precipitated 
silver salt is filtered, washed with hot, dilute nitric acid, dried and weighed as Agl. 
The filtrate containing the silver, formerly with the chlorine and bromine, is 
treated with potassiimi iodide in sufficient amount completely to precipitate 
the silver as Agl. This is filtered, washed and weighed. From the three weights 
the chlorine, bromine and iodine can be easily calculated. 

Note. Bekk claims an accuracy within less than 0.15%. 

EVALUATION OF BLEACHING POWDER, CHLORIDE OF 

LIME, FOR AVAILABLE CHLORINE 

When chloride of lime is treated with water, it is resolved into calcium 
hypochlorite, Ca(0Cl)2, and calcium chloride, CaClj. The calcium hypoclilorite 
constitutes the bleaching agent. The technical analysis is confined to the 
determination of available chlorine, which is expressed as percentage by weight 
of the bleaching powder.* 

Procedure. Ten grams of the sample are washed into a mortar and ground 
with water, the residue allowed to settle and the supernatant liquor poured into 
a liter flask. The residue is repeatedly ground and extracted with water until 
the whole of the chloride is transferred to the flask. The combined extracts 
are made up to 1000 cc. 

To 50-cc. portions (0.5 gram) of the solution, 3 to 4 grams of solid potassium 
iodide and 100 cc. of water are added and the solution acidified with acetic 
acid. Iodine equivalent to the available chlorine Ls liberated. This is titrated 
with N/10 arsenious acid.' 

One cc. N/10 arsenious acid =0.003546 gram CI. This multiplied by 200 = %C1. 

Uulius Bckk, Chem. Ztg.. 89, 405-6 (1915). C. A., 9, 2042, (1915). 

* In France the strcngtn is given in Gay-Lussac degrees, e.g., liters of gas 
evolved by 2 kilograms of bleaching powder, 0° (J. and 760 mm. 100° = 31 .78% CI. 

' The standard arsenious acid is made by dissolving 4.95 grams of pure AstOs 
together with 20 grams of sodium bicarbonate in 50 cc. of warm water. When dis- 
solved the solution is made up to 1 liter. 



CHLORINE 131 

Note. In the analysis of compounds containing hypochlorites and chlorides, 
the conversion of hypochlorites to chlorides by heating with hydrogen peroxide is 
a great convenience. 

For instances in the analysis of bleach liquors, washes, etc., the (OCl) and CI 
may be very easily and quickly determined by titrating an aliquot with As^Os and then 
a similar aliquot with AgXOs after converting all the OCl to CI by warming with 
HiO,. 

It is also a convenience in getting rid of OCl as, for instance, in the deter- 
mination of COs in bleaching powder, which is often of great importance. It is 
preferable to the use of ammonia, which is always liable to suspicion of having taken 
up a little COt, and there is no danger of NH4CI fumes which are sometimes a nuisance. 



CHROMIUM 

Wilfred W. Scott 

Cr, at.wt. 53.0; sp.gr. 6.92; m.p. 1530''; h.p. 2200"" C; oxides, CrOs; 

Cr^O,, CrOs. 

DETECTION 

Chromium is precipitated by hydrogen sulphide and ammonium hydroxide 
as bluish-green, Cr(OH)i, along with the hydroxides of iron and aluminum 
(members of previous groups having been removed). The chromic compound 
is oxidized to chromate by action of chlorine, bromine, sodium peroxide, or 
hydrogen peroxide added to the substance containing an excess of caustic alkali. 
The chromate dissolves and is thus separated from iron, which remains insol- 
uble as Fe(OH)i. The alkali chromates color the solution yellow. 

Barium acetate or chloride added to a neutral or slightly acetic acid 
solution of a chromate precipitates yellow barium chromate, BaCrOi. Addition 
of ammonium acetate to neutralize any free inorganic acid aids the reaction. 

Lead acetate produces a yellow precipitate with chromates, in neutral 
or acetic acid solutions. 

Mercurous nitrate or silver nitrate gives red precipitates with chromates. 

Hydrogen peroxide added to a chromate and heated with an acid, such 
as sulphuric, nitric, or hydrochloric, will form a greenish-blue colored solution. 
Chromates are reduced by hydrogen peroxide in acid solution, the action being 
reversed in alkaline solution. 

Reducing agents, hydrogen sulphide, sulphurous acid, ferrous salts, 
alcohol form green chromic salts when added to chromates in acid solution. 

Ether shaken with a chromate to which nitric acid and hydrogen peroxide 
are added, is colored a transient blue. Oxygen is given off as the color fades. 

HCr04+3HNO, =Cr(NO,),+2H,0-F02. 

Diphenyl carbazide test. To 5 cc. of the solution containing chromium as 
chromate, 2 drops of hydrochloric or acetic acid are added, and 1 drop of an acetic 
acid solution of diphenyl carbazide (0.2 gram CO (NH-NH-C6H6)2 is dissolved 
in 5 cc. glacial acetic acid and diluted to 20 cc. with ethyl alcohol). A violet 
pink color is produced in presence of a chromate. Less than 0.0000001 gram 
chromium may be detected. 

Chromic salts are bluish green; chromic acid is red; chromates, yellow; 
bichromates, red; chrome alum, violet. 

The powdered mineral, containing chromium, when fused with sodium 
carbonate and nitrate, produces a yellow colored mass. 

ESTIMATION 

Among the substances in which chromium is determined are the following: 
Chrome iron or chromite, CrjOj-FeOMgO; crocoisite, PbCr04; slags; chromic 
oxide, chrome green, in pigments; chromates and dichromates; chrome steel 
and ferro-chrome. 

132 



CHROMIUM 133 



Preparation and Solution of the Sample 

Although powdered metallic chromium is soluble in dilute hydrochloric or 
sulphuric acid, it is only slightly soluble in dilute or concentrated nitric acid. 
It is practically insoluble in aqua regia and in concentrated sulphuric acid. 
Chrome iron ore is difficult to dissolve. It is important to have the material 
in finely powdered form to effect a rapid and complete solution of the sample. 
An agate mortar may be used to advantage in the final pulverizing of the 
substance. 

General Procedures for Decomposition of Refractory Materials Con- 
taining Chromium. The following fluxes may be used: 

A. Fusion with KHSO4 and extraction with hot dilute HCl. The residue 
fused with NajCOi and KClOa, 3 : 1, or fusion with soda lime and KClOj, 3:1. 

B. Fusion with NaHSO* and NaF, 2:1. 

C. Fusion with magnesia or lime and sodium or potassium carbonates, 
4 :1. 

E. Fusion with Na,0,, or NaOH and KNO,, or NaOH and NajOa. Nickel, 
iron, copper, or silver crucibles should be used for E, Platinum may be used 
for A, B, or (7. 

Special Procedures. Materials High in Silica. The finely ground sample, 
1 to 5 grams, is placed in a platinum dish and mixed with 2 to 5 cc. concentrated 
sulphiiric acid (1.84), and 10 to 50 cc. of strong hydrofluoric acid added. The 
solution is evaporated to small volume on the stcom bath and to SO3 fumes on 
the hot plate. Sodium carbonate is added in sufficient amount to react with 
the free acid, and then an excess of 5 to 10 grams added and the mixture heated 
to fusion and kept in molten condition for half an hour. From time to time a 
cr3r8tal of potassium nitrate is added to the center of the molten mass until 
1 to 2 grams are added. {Caution, Platinum is attacked by KNO», hence 
avoid adding a large amount at any one time.) Chromium and aluminum go 
into solution in the flux, but iron is thrown out as Fe(OH)j. The cooled fusion 
is extracted with hot water and filtered from the iron residue. Chromium 
is in solution together with aluminum. If much iron is present it should be dis- 
dissolved in a little hydrochloric acid and the solution poured into, boiling 10% 
solution of potassium hydroxide, the cooled solution +Fe(OH) 3 precipitate is 
treated with hydrogen peroxide or sodium peroxide to oxidize any chromium 
that may have been occluded by the iron in the first precipitate. The mixture 
IS again filtered and the combined filtrates examined for chromium. 

Sodium Peroxide Fusion. Chrome Iron Ores. One to two grams of finely 
pulverized ore are placed in a nickel or iron crucible of 50 to 75 cc. capacity 
and mixed with 5 to 10 grams of yellow sodimn peroxide. (Fresh peroxide is 
best). The mass is gently heated over a Bunsen burner until it melts. The 
fusion is kept at a low red heat for about fifteen minutes. About 5 grams more 
of the NaiOs are added and the fusion heated for about ten minutes more. 
The cooled fusion is dissolved in a casserole with 100 cc. to 150 cc. of water, more 
peroxide being added to this solution if it appears purple. The excess of peroxide 
is decon^posed by boiling the solution, and to the caustic solution free from per- 
oxide is added 10 to 15 grams of ammonium carbonate or a sufficient quantity of 
the salt to neutraUze four-fifths of the sodium hydroxide present in the solu- 
tion, as the strong caustic would otherwise dissolve the filter. The solution 
is now filtered. The insoluble matter is treated on the filter with dilute sulphuric 



134 CHROMIUM 

acid, 1:4. If a portion remains insoluble, it is an indication of incomplete 
decomposition of the ore, and this residue is again fused with peroxide and 
treated as above. The combined filtrates contain the chromium. 

Since chromates are reduced in presence of free acid and peroxide, the latter 
should be expelled before making the solution acid.* 

If the chromate is to be precipitated as BaCrOi or PbCr04, the solution 
should be acidified with hydrochloric acid. If the reduced solution is to be 
titrated with potassium permanganate, it is best to use sulphuric acid in neutral- 
izing the caustic solution. Further directions will be given under the method 
chosen. 

Method for Solution of Iron and Steel. Three to five grams of steel are 
boiled for about ten to fifteen minutes with 50 cc. of strong hydrochloric acid 
and about 150 cc. concentrated nitric acid added and the boiling continued 
until the hydrochloric acid is expelled, brown fumes and the odor of CI no longer 
being evident. Ten grams of potassium chlorate are now added to the cooled 
solution, a few cr>^stals at a time, the solution then replaced on the hot pbte 
and boiled down to about 50 cc. to decompose the chlorate. The solution is 
diluted to 150 cc, and if an appreciable amount of manganese dioxide has 
separated out, it is decomposed by the addition of a few drops of hydrochloric 
acid. The chlorine is expelled by boiling and the solution cooled. Chromium 
is determined in this solution by the ferrous ammonium sulphate method. 

SEPARATIONS 

Chromium, Iron, and Aluminum. If chromium has been fused with sodium 
peroxide or carbonate containing a little potassium nitrate, and the fusion 
extracted with boiling water, most of the chromium goes into solution as a 
chromate, together with alumina, but aome of the chromium is occluded by Fe(OH)i. 
If the amount of the iron precipitate is appreciable, and warrants the recovery 
of occluded chromium, it is dissolved in hydrochloric acid and the iron repre- 
cipitated by pouring into a solution of strong sodium hydroxide. Before filtering 
off the iron hydroxide, a little HjOs is added to oxidize the CraOs, if accidentally 
present, and the solution boiled and filtered. The combined filtrates will con- 
tain all of the chromium and aluminum. 

If chromium is present as a chromic salt, instead of a chromate, it is oxidized 
to the higher form, by adding peroxide (HjOa or NaaOj) to the alkaline solu- 
tion. Bromine added to this solution or chlorine gas passed in will accomplish 
complete oxidation.' It must be remembered that in acid solutions hydrogen 
peroxide, sodium peroxide, or nitrites will cause reduction of chromates to 
chromic salts (exception, see method for solution of steel), so that these should 
be boiled out of the alkaline solution before making decidedly acid with hydro- 
chloric or sulphuric acids. Since these are difficult, if not impossible, to com- 
pletely expel from an alkaline solution, after boiling the strongly alkaline solu- 
tion, dilute sulphuric acid is added until the solution acquires a permanent 
brown color (nearly acid), acid potassium sulphate, KHSO4, is added, and 

* See Separations. 

* Br may be added and then NaOH to oxidize Cr and precipitate Fe(OH)s. 
Chromic oxide and moat of ita com|)ound8, except chrome iron stone, may be 

decomposed bv cone. HN()|-f-KC10i (added in small portions). M. Gr6ger, Zeitsch. 
anorg. Chem./81, 233-242, 1913. 



CHROMIUM 135 

the boiling continued.^ This will decompose the bromates and expel bromine, 
etc., but will not cause the reduction of the chromate, as would a strong acid 
solution. 

Separation of Chromium from Aluminum. This separation is necessary if 
chromium is to be precipitated as CrCOH),. The sodium chromate and alurai- 
nate solutions are made slightly acid with nitric acid and then faintly alkaline 
with anunonium hydroxide, Al(OH)s is precipitated and chromium remains 
in solution as a chromate. 



GRAVIMETRIC METHODS FOR THE DETERMINATION OF 

CHROMIUM 

Precipitation of Chromic Hydroxide and Ignition to Cr203^ 

Chromium present as a chromic salt in solution, free from iron and aluminum 
or elements precipitated as hydroxides, is thrown out of solution by NH4OH 
as Cr(OH)i, the precipitate ignited to the oxide, CriOa,' and so weighed. The 
presence of hydrochloric acid or sulphuric acid does not interfere. 

Reductioii. U the chromium is already present as the chromic salt, free 
from iron and alumina, it may be precipitated directly as the hydroxide by 
addition of ammonia; otherwise, if present as the chromate, as is the case when 
a separation from iron and alumina has been necessary, and in cases where 
the chromium has been brought into solution by fusion with an oxidizing 
reagent, reduction is necessary. This is accomplished by passing SOi or 
HjS into the slightly acid solution of the chromate, or by adding alcohol 
to the hydrochloric acid solution and boiling until the solution appears a 
deep grass green. Twenty cc. of alcohol for every 0.1 gram of Or has been 
found to be ample for this reduction. The SO2 or H2S should be expelled 
from solution by boiling, in case either has been used for reduction of the 
chromate. 

Precipitation. Ammonium hydroxide or ammonium sulphide are added 
in slight excess and the solution boiled for about ten minutes. The solution 
should be slightly alkaline (litmus), otherwise a few drops of ammonia should 
be added, but not a large excess; the solution will then settle out clear. A 
cloudy solution results from prolonged boiling when the solution has become 
acid; on the other hand, a large excess of ammonia will prevent complete pre- 
cipitation of chromium and the filtrate will be colored pink or violet. The 
chromic hydroxide is filtered off on S and S 689 filter paper. Since the precipi- 
tate is apt to be gelatinous it is advisable to wash two or three times by decanta- 
tion and several times on the paper. The well-drained precipitate and filter is 
ignited wet in a porcelain or platinum crucible, first over a low flame until the 
paper has been charred, then over a strong gas flame for about thirty minutes, 

^KHSO* will not cause reduction of chroraates. A. Kiu-tenacker, Zeitsch. anal. 
Chem., 62, 401-407, 1913. The Analyst, 38, 449, page 387. 

* It is advisable to take such a weight of sample that the ignited CrjOa does not 
exceed 0.5 gram in weight. 

•CrjOj, mol.wt.y 152; 8p.gr.j 5.04; m.p., 2059°; insol. in HaO, slightly sol. in acids, 
dark green hexagonal. 



136 CHROMIUM 

and finally a blast heat for five minutes. The green residue is weighed as 

Cr2O,X0.6846=Cr. 

Determination of Chromium as Barium Chromate^ 

Chromium, present as a chromate, is precipitated from a neutral or faintly 
fkcetic acid solution of an alkali chromate by addition of barium acetate or 
chloride. The BaCrOi is gently ignited and weighed. The solution should be 
free from sulphuric acid or sulphates. 

Procedure. The alkali chromate solution is neutralized with nitric acid 
or ammonia as the case may require, precautions for avoiding reduction 
having been observed as indicated under Preparation and Solution of the 
Sample. 10 cc. of \ N. BaClj or BaCCsHjOi), (approx. 10% sol.) are added 
to the boiling solution for each 0.1 gram of chromium present. The reagent 
should be added in a fine stream or drop by drop to prevent occlusion of the 
reagent by the precipitate. The precipitated chromate is allowed to settle 
on the steam bath for two or three hours and then filtered into a weighed Gooch 
crucible and washed with 10% alcohol solution. The precipitate is dried for 
an hour in the oven, then placed in an asbestos ring suspended in a large 
crucible with cover and thus heated over a low flame, gradually increasing the 
heat until the outer crucible becomes a dull red. The cover is removed and 
the heating continued for five minutes, or until the precipitate app)ears a uniform 
yellow throughout. High heating should be avoided. The cooled residue is 
weighed as BaCr04. 

BaCrO* X 0.2055 =Cr, 

BaCr04X 0.3002 =Cr,0,. 

BaCrOiX 0.7666 =KjCr04. 

BaCr04X 0.5807 ^RCrA. 

Notes. If the precipitate on the sides of the crucible appears green, it is ignited 

until the green color disappears. 

If sulphates are present, BaSOi will be precipitated, hence this method could 
not be used. In this case either reduction to the chromic salt and preci|>itation of 
chromium as Cr(OH)j or a volumetric procedure should he followed. 

Oxidize chromium with an excess of hydrogen peroxide in alkaline solution, reduce 
in acid solution with ferrous sulphate and titrate with perman^^anate. Decomposition 
of hydrogen peroxide is accelerated by heat and by presence of sodium sulphate or 
ferric salt«. Salts of nickel cobalt, or manganese, decompose HiOj energetically and 
lower results are obtained. F. Bourin and A. Senechal. Compt. lend., 167, 1528-31. 

^ If the filtrate appears 3'ellow, chiomate is indicated, the solution should he reduced 
and the chromium precipitated as Cr(OFI)j. If the filtrate is f)ink, it should be }>oiled 
until it appears green and Cr(C>H)3 precipitates. These precii)itate3 should be included 
in the above calculation for chromium. 

«BaCr(J4, mo/.irr, 253.47; sp.gr., 4.498; solubility per 100 cc. H2O, 0.00038 »•" and 
0.0043 hot. Soluble in HCl and in HNO3; yellow rhombic i)lates. 



CHROMIUM 137 

VOLUMETRIC METHODS FOR THE DETERMINATION 

OF CHROMIUM 

Potassium Iodide Method for Determination of Chromium 

Chromium present as a chromate is reduced in acid solution by addition 
of potassimn iodide and the liberated iodine titrated by standard sodium thio- 
sulphate. The method depends upon the following reactions: 

(a) 2CrO,+6KI =Cr,0,+3K20+6L 
(6) l2+2Na2S,0,=2NaI+Xa,S40.. 

The presence of large quantities of Ca, Br, Sr, Mg, Zn, Cd, Al, Ni, Co, H2SO4, 
HCl, does not interfere.* 

Procedure. The alkali chromate solution containing not over 0.17 gram 
Cr * and free from FeaOj, is made nearly acid with H2SO4, boiled with 20 cc. of 
30% potassiimi acid sulphate to decompose bromates or expel Br, CI, or H2O2 
as the case may require, more KHSO4 being added if necessary. If the solution 
is not acid it is made so with sulphuric acid and 5 cc. of the acid per 100 C3. of 
solution is added in excess.' About 2 grams of solid potassium iodide arc added 
and, after five minutes, the liberated iodine is titrated with N/10 Na2S203 
solution. When the green color of the reduced chromate begins to predominate 
over the free iodine color (brownish red) a little starch solution is added and 
the titration with the thiosulphate continued until the blue color of the starch 
compound is just destroyed, care being taken not to confuse the green color of 
the reduced chromium with the blue of the starch. 

One cc. of N/10 NatSaO,* =0.001733 gram Cr. 

Determination of Chromium by Reduction of the Chromate 

with Ferrous Salts 

The procedure may be used for the determination of chromium in presence 
of ferric iron and alumina. Hydrochloric or sulphuric acids do not interfere. 
If hydrochloric acid is present in solution, the K2Cr207 back titration should be 
made. In presence of H2SO4 either KMn04 or K2Cr207 titrations may be 
made. The method depends upon the reduction of soluble chromatcs by ferrous 
salts, the excess being determined by titration. 

Reactions, a. 2Cr20,+6FeO+x«FeO=Cr203+3Fe203+X5FeO. 

b. xsFeO is oxidized by standard oxidizing reagent to Fe20|. 

iM. Gr6ger, Zeit. anal. Chem., 81, 233-242, 1913. 

*If desired, btronger solution of titration reagentd may be used, and consequently 
a larger sample taken. A nonnal sol. of Na^SiOj may be used to advantage with 1 
gram samples of chromium salts or hydrates, where Cr exceeds 10%. 

» A. Kurtenacker, Zeit. anal. Chem., 52, 401^07. 1913. 

Sutton recommends for every 0.5 gram KsCreOy present to add .5 gram KI and 1 .8 
gram H2SO4 per 100 cc. of solution. If more K2Cr207 is present, increase the KI and 
H2SO4, but not the water. 

*It desired, a normal solution of thio sulphate may be used with one gram sample 
of chromium salts or hydroxides, when the chromium present exceeds 10 per cent. 



138 CHROMIUM 

Procedure. Redvciion, The sample, containing not over 0. 17 gram chromium 
present as a chromate, is boiled to expel oxidizing reagents according to the 
method described under the potassium iodide procedure for chromimn. The solu- 
tion is made acid, if not already so, and about 5 cc. cone. H2SO4 per 100 cc. 
of solution, added in excess. Tenth normal ferrous ammonium sulphate solu- 
tion containing free sulphuric acid is added until the solution changes from 
yellow through olive green to deep grass green. For every 0.1 gram of chro- 
mium about 65 to 70 cc. of N/10 ferrous salt solution should be added. After 
five minutes, the excess of this reducing reagent is titrated either with perman- 
ganate or with dichromate as directed below. 

Potassium Permanganate Titration. To be used in presence of free sulphuric 
acid, free hydrochloric acid being absent. 

Tenth-normal potassiiun permanganate solution is run into the reduced 
chromate until the green color gives place to a violet tinge. At the end-point 
the solution appears to darken slightly. A little practice enables one to get this 
with accuracy. A slight excess of permanganate gives the solution a pinkish 
color, readily distinguishable in the green. Addition of 3 to 4 cc. syrupy phos- 
phoric acid gives a sharper end-point. The color should hold one minute. 

Potassium Dichromate Titration. N/10 KzCrsO?^ is run into the solution 
until a drop of the sample placed on a white glazed surface with a drop of potas- 
siiun ferricyanide reagent no longer gives a blue color. 

Calculation. From the total ferrous ammoniiun sulphate added, subtract 
the cc. of back titration (the reagents being exactly N/10), the difference gives 
the cc. of ferrous salt required for chromium reduction. If reagents are not 
N/10, multiply cc. titrations by factor converting to N/10. 
Cc. ferrous ammonium sulphate X 0.001 733 =Cr. 

CraOa+SO =Cr,Oe. .*. Cr = 1^0 or =3H; hence § mol. wt. Cr per liter =N sol. 

Determination of Small Amounts of Chromium 2 

Advantage may be taken of the color produced by chroma tes in solution 
in determining small amounts, the depth of color depending upon the amount 
of chromate in solution. The method possesses the usual disadvantage of color- 
imetric procedures in that there is always room for doubt as to whether the 
element sought is entirely responsible for the color of the solution. 

Procedure. The solution containing the sample is nearly neutralized with 
sodium carbonate, the reagent being added until a slight cloudiness results. 
The solution is now cleared with a few drops of sulphuric acid, and then suf- 
ficient excess of a strong solution of sodium thiosulphate added to precipitate 
aluminum, chromium, manganese, etc. The precipitate is filtered off, dis- 
solved in the least amount of dilute nitric acid, then filtered from the precipitated 
sulphur and diluted to 300 to 400 cc. Chromium is now oxidized by adding 
10 cc. of 0.2% silver nitrate solution, about 10 grams each of ammonium nitrate 
and persulphate. After boiling for about twenty minutes, sufficient hydro- 
chloric acid is added to decompose any pcnnanganate present and to precip- 

*If desired, a larger sample may be taken and N/5 or N solutions used in titra- 
tion. It is advisable to titrate chromium salts, e.g., over 1.0% Cr, with normal solu- 
tions, so that one gram sample may be taken for analysis. 

» M. Dittrich, Zeitsch. anorg. Chem., 80, 171-174, 1913. 



CHROMIUM 139 

itate the silvei*, and a few cc. added in excess. The solution is again boiled 
for about ten minutes and then filtered. The filtrate is treated with a little 
sodium phosphate to repress the color of traces of iron that may be present 
and made to a definite volume. 

The solution may now be compared with a standard solution containing 
the same amounts of acids, manganese, alumina, etc., as are present in the 
sample, tenth normal potassium dichromate being run into this standard solu- 
tion until its color matches that of the sample. The burette reading is taken 
and the chromium calculated. 

One cc. of N/10 K^Crsa =0.00173 gram Cr. 

Notes. Prolonged boiling after addition of hydrochloric acid to the solution of 
the chromate will cause its reduction. A green tint usually indicates that the chro- 
mate has been reduced. 

The test may be carried on in the presence of sulphiu'ic, hydrochloric, phosphoric, 
hydrofluoric, and nitric acids. Alumina, manganese, and small amounts of iron 
do not interfere. 

Organic matter should be destroyed by either calcining the sample or by oxida- 
tion by taking to fumes with sulphiu'ic acid. The presence of this prevents pre- 
cipitation of chromiiun. 



COBALT 

W. L. Savell^ 

COt aUwt. 58.97; sp.gr. 8.7918; m,p. 1478^; b.p. unkru>wn; Oxides, CoiO^* 

CosOs, CoO, CoOa. 

DETECTION 

After the removal of the elements precipitated by hydrogen sulphide from 
acid solution, a little nitric acid is added to the solution to oxidize to the ferric 
state any ferrous salts which may be present, and ammonia is added until its 
odor is distinctly perceptible, to precipitate iron, aluminum and chromium.* 
This precipitate is removed by filtration and hydrogen sulphide passed through 
the ammoniacal solution to precipitate cobalt, nickel, manganese and zinc. After 
collecting this precipitate it is washed thoroughly with cold hydrochloric acid of 
approximately 1.035 specific gravity, to remove manganese and zinc. A small 
quantity of the residue is fused with borax in the loop of a platinum wire. A 
blue color in the cold bead indicates cobalt. This test is masked in the presence 
of large quantities of nickel. In this case the residue is dissolved in hydrochloric 
acid to which a few drops of nitric acid have been added and the solution evap- 
orated to dr\'ness. The residue is redissolved in water, acidified with hydrochloric 
acid and the cobalt precipitated with a hot solution of nitroso-beta-naphthol 
in 50% acetic acid. A brick red precipitate indicates cobalt. 

Potassium sulphocyanate, KCNS, produces a red color with cobalt. Alcohol 
and ether are added to this solution and shaken. The ether layer is colored 
blue by cobalt. If iron is present a solution of sodium thio-sulphate, NajSjOa, 
is added until the red color disappears, the solution filtered and then treated with 
the alcohol-ether mixture. 

Potassium Nitrite, KNOj, added to a neutral or slightly acid solution con- 
taining acetic acid, will precipitate cobalt as a yellow complex nitrite having 
the formula KaCo(N02)6. 

A solution of dicyandiamidine sulphate and sodium hydroxide added to a 
cobalt solution to which ammonia has been added until the odor is distinctly 
discernible, and containing from 10 to 20 cc. of 10% sugar solution, will change 
the color of the solution to red or reddish violet. If large quantities of nickel 
are present the color will be yellow or reddish yellow, after which the nickel will 
separate out in brilHant crj'stals, leaving the cobalt in solution, coloring it as 
described above. 

A concentrated solution of ommonium sulphocifannie added to a cobaltous 
solution colors it blue. On dilution this becomes pink. Amyl alcohol or a mix- 
ture of amyl alcohol and ether 1:1, added to this and shaken, extracts this 
blue compound. Iron sulphocyanate, Fe(CNS),, likewise colors the ether- 

* Research Chemist, Doloro Mining and Reduction Company, Doloro, Ontario. 

* If a relatively large anunint of iron is present the basic acetate method of separa- 
tion is necessary, as iron occludes cobalt. 

140 



COBALT 141 

alcohol extract red, which may mask the cobalt blue. By addition of sodium 
carbonate solution ferric hydroxide precipitates, while the cobalt color will remain 
after this treatment. 

ESTIMATION 

Cobalt is usually estimated as metal; either reduced by hydrogen from the 
ignited oxide or reduced by electrolysis from an ammoniacal solution of its 
salts. Sometimes, however, it is estimated as oxide; usually as C03O4. The 
reduction of the oxide by hydrogen may be carried out in conjunction with any 
process giving an oxide, hydroxide, carbonate, nitrate, chloride or an organic 
compound, as a final product. 

The reduction of the metal, in solution, by electrolysis, must be accomplished 
in a strongly ammoniacal solution free from copper and nickel, as these metals 
are deposited with the cobalt on the cathode. When desirable the copper and 
nickel may be estimated after the electrolysis by dissolving the deposit from the 
cathode and proceeding in the usual manner. 

Preparation and Solution of the Sample 

General Procedure for Ores. The ores containing cobalt vary so widely 
in their chemical nature that it is difficult to lay down a method for treating all 
ores. However, as the principal ores contain the cobalt as a sulphide or arsenide 
the same general methods may be used in the majority of cases. In all cases it 
is necessary to prepare the sample for treatment by grinding finely. Usually 
either of the above ores may be brought into solution by heating with strong 
nitric acid or a mixture of nitric and hydrochloric acids, except silver-bearing 
ores, which may usually be dissolved in a mixture of nitric and sulphuric acids. 

While it is desirable to use no more acid than is necessary to bring the sam- 
ple into solution, an excess will not interfere, as it may be driven off by evapora- 
tion and in the event of determining the cobalt electrol3rtically it is essential 
that the solution be free from nitric acid, so that this evaporation becomes 
part of the procedure. 

In the case of especially refractory ores or oxides of cobalt or nickel, a fusion 
with potassium bisulphate will usually be found sufficient as a preliminary treat- 
ment to enable it to be brought into solution. Under certain conditions, how- 
ever, it has been found necessary to fuse the ore with sodium peroxide in a silver 
crucible, dissolving the cobalt oxide formed in hydrochloric acid. In some- 
what less refractory ores of a silicious nature a preliminary fusion with a mixture 
of sodium carbonate and potassium carbonate with subsequent solution in 
hydrochloric acid or sulphuric acid, if the ore is a silver-bearing one, will be 
found satisfactory. 

Cobalt Oxides. Cobalt oxide, gray or black, may be fused with potassium 
bisulphate, and the melt leached with water; or they may be treated with sul- 
phuric acid, in which they dissolve slowly; or with hydrochloric acid, in which 
they dissolve more rapidly. 

Metallic Cobalt, Nickel and Cobalt Alloys. Metallic cobalt dissolves 
readily in nitric acid, as do nickel and the ordinary cobalt alloys. There are 
some alloys of cobalt, however, which require fusion with sodium peroxide be- 
fore they become amenable to further treatment. Among these are certain 
cobalt-chromium alloys. 



142 COBALT 



SEPARATIONS 

Separation of the Ammonium Sulphide Group Containing Cobalt from the 
Hydrogen Sulphide Group — Mercury, Lead, Bismuth, Copper, Cadmium 
Arsenic, Antimony, Tin, Gold, Molybdenum, etc. 

Hydrogen sulphide passed into a hydrochloric acid solution containing 
from 5 to 7 cc. of concentrated hydrochloric acid per 100 cc. of solution, pre- 
cipitates only the members of that gi'oup and silver, whereas the members of the 
subsequent groups remain in solution. If the solution is too acid, lead and 
cadmium are not completely precipitated. 

Separation of the Ammonium Sulphide Group from the Alkaline Earths 
and Alkalies. Anunoniiun sulphide, free from carbonate, ac^ded to a neutral 
solution containing the above elements in the presence of anmxonium chloride, 
precipitates only the members of this group; the alkaUne earths metals, mag- 
nesium and the alkaUes remain in solution. A second precipitation should be 
made if large quantities of the alkaline earths or alkalies are present. 

Separation of Cobalt and Nickel from Manganese. The solution of the 
chlorides or sulphates of cobalt or nickel is treated with an excess of sodium 
carbonate and then made strongly acid with acetic acid. About 5 grams of 
sodium acetate for each gram of cobalt or nickel present is now added, the solution 
diluted to 200 cc. and heated to about 80° C. and saturated with hydrogen sul- 
phide. Cobalt and nickel are precipitated as sulphides and the manganese 
remains in solution. The filtrate is concentrated, and colorless anunonium 
sulphide added when the cobalt and nickel that may have passed in solution 
from the hydrogen sulphide treatment, will be precipitated. The treatment 
should be repeated with the second filtrate to ensure complete precipitation of 
the cobalt and nickel. 

Separation of Cobalt from NickeL Among a number of methods for effect- 
ing this separation the following give good results: 

A, Nickel is removed from the solution by precipitation with dunethyl- 
glyoxime. The details of the procedure may be found in the gravimetric methods 
for the determination of nickel. Cobalt remains in solution. 

B, Cobalt is precipitated by nitroso-beta-naphthol, leaving nickel in solu- 
tion. Details of the procedure are given under gravimetric methods for deter- 
mination of cobalt. 

C, Cobalt is precipitated as potassium cobalti-nitrite, nickel remaming in 
solution. Details of the procedure are given under gra\'imetric methods for 
the determination of cobalt. 

Separation of Cobalt from Zinc. Zinc is precipitated from weak acetic or 
formic acid solution by hydrogen sulphide as zinc sulphide. Cobalt, nickel 
and manganese remain in solution. The details of the procedure are given under 
the methods of determination of zinc. 



COBALT 143 

GRAVIMETRIC METHODS FOR THE DETERMINATION 

OF COBALT 

Precipitation of Cobalt by Potassium Nitrite 

Cobalt may be precipitated from a solution made slightly acid with an excess of 
acetic acid by adding a hot solution of potassium nitrite. The cobalt is precipitated 
as potassium cobalti-nitrite, KsCo(NOa)«, very completely, after standing for a 
period of six hours in a warm place. This method has the advantage of making 
possible the separation of cobalt from nickel and iron, although it has the one 
disadvantage, for commercial purposes, of requiring a long time to complete the 
determination. 

Procedure. After bringing the material into solution and separating the 
silica and members of the first and second groups in the usual manner, the 
solution is boiled to eliminate hydrogen sulphide. Oxidize the iron present 
with a little hydrogen peroxide and evaporate the solution to a syrup. Take 
up in a little water and neutralize with a practically saturated solution of sodium 
carbonate. Render slightly acid with acetic acid and add an excess of 1 : 1 
acetic acid. Heat to almost boiling and add solution of 50% potassium nitrite 
containing 100 cc. of glacial acetic acid per liter, also heated to nearly boihng. 
This solution should be added slowly to the solution of the sample which should 
be agitated, preferably by rotating gently while the addition is being made. 
The sides of the beaker should be washed down with a 1% solution of potassium 
nitrite containing 1 cc. of glacial acetic acid per liter. Allow to stand for at 
least six hours and if possible overnight. Filter through thick Swedish filter 
paper without previous wetting. As this precipitate shows a very decided 
tendency to creep, considerable care is required to keep it well down in the apex 
of the filter-paper cone. Wash about ten times with the warm nitrite solution 
mentioned above. Transfer to a beaker by removing the filter paper from the 
funnel and opening it into the beaker with the outside of the paper against the 
glass. This leaves it in a convenient position for washing. The bulk of the 
precipitate is washed off with 10 cc. of 1 : 1 sulphuric acid, heated to about 80® C. 
This should leave only a slight film of precipitate on the paper. Keep the solution 
in the beaker at about 80® C. to assist in dissolving the precipitate and wash 
the paper with the hot sulphuric acid solution five times, using about 10 cc. each 
time. Gradually withdraw the filter paper from the top of the beaker during the 
washing operation. Give the paper a final wash with hot water and squeeze the 
last drops from it into the beaker. Evaporate and allow to fume strongly for 
ten minutes. Set the beaker in a cooling trough and add water until the volume 
is about doubled. Neutralize and make slightly ammoniacal and then add 
an excess of 50 cc. of strong ammonia and electrolyze as described under Elec- 
trolysis in Reduction of Cobalt by Electrolysis, page 144. 

Precipitation of Cobalt by Nitroso-beta-NaphthoI ^ 

Nitroeo-beta-naphthol, CioH«(NOH), added to a hydrochloric acid solution of 
cobalt, precipitates cobalti-nitroso-beta-naphthol, CoCCioHeOCNO))^; nickel, if 
present, remains, in solution. The method is especially suitable for the deter- 
mination of small amounts of cobalt in the presence of comparatively large 

» Burgess, Z. Angew., 1896, 696. 



144 COBALT 

amounts of nickel. The cobalt precipitate is voluminous, so that the sample 
taken for the determination should not contain over 0.1 gram of cobalt. The 
reagent will also precipitate copper and iron completely from solution, and 
silver, bismuth, chromium and tin partially; but mercury, lead, cadmium, arsenic, 
antimony, aluminum, manganese, nickel, glucinum, calcium and magnesium 
remain in solution. 

Procedure. To the solution containing the cobalt is added a freshly prepared 
hot solution of nitroso-betarnaphthol, in 50% acetic acid, as long as a precipi- 
tate is produced. Aftei* allowing it to settle, more of the reagent is added to 
insure complete precipitation of the cobalt. The compound is allowed to 
settle for two of three hours, the clear solution decanted through a filter and the 
precipitate washed by decantation with cold water, then with warm 12% hydro- 
chloric acid solution to remove the nickel, and finally with hot water until free of 
acid. 

The brick-red precipitate is dried, then ignited in a weighed platinum cru- 
cible (Rose crucible), first over a low flame and finally at a white heat, the 
crucible being covered by a platinum cover (Rose crucible type) with a platinum 
tube, through which is passed a slow current of oxygen. The residue is weighed 
as C03O4. The oxide may be reduced in a current of hydrogen and weighed as 
metallic cobalt. Ignited in the presence of C0» the oxide CoO is formed. 

Precipitation of Cobalt by Electrolysis^ 

Metallic cobalt is readily deposited from an ammoniacal solution of the 
sulphate, but in the presence of copper and nickel these are also completely pre- 
cipitated on the cathode; so, in case it is desired to determine the cobalt alone 
it is necessary to separate these metals from the solution before electrolysis or to 
determine them separately after electrolysis in a solution of the metallic deposit. 
In practice the copper is usually separated before electrolysis and the nickel, if 
determined separately, is estimated afterward by one of the methods given under 
Nickel, the cathode deposit being dissolved for this purpose. 

Procedure. After preparation and solution of the sample the usual sepa- 
rations with hydrogen sulphide in acid solution are made if necessary. In 
most cases it is necessary to pass hydrogen sulphide through the warmed solu- 
tion for at least one hour to insure the complete precipitation of arsenic. Filter 
and boil to expel hydrogen sulphide. Add 5 cc. hydrogen peroxide to insure 
oxidation of iron compounds to ferric state and add ammonium hydroxide until 
slightly alkaline to litmus. Filter off ferric hydroxide and wash with water 
containing a small quantity of ammonium hydroxide. Redissolve and repre- 
cipitate this ferric hydroxide in the above manner, using a little hydrogen per- 
oxide in each instance, until the last traces of cobalt have been removed from it, 
keeping the filtrates, which should be as small as possible, to add to the main 
filtrate. If much iron is present this is best removed as the basic acetate. 

Electrolysis. If the treatment of the iron precipitate has made a large 
volume of solution this may be reduced by evaporation, after which 50 cc. of 
strong ammonia are added and the solution electrolyzcd, using direct current of 
2 volts and 0.5 ampere per square decimeter. The electrodes should be 
platinum, the anode a spiral wire and the cathode either a hollow cylinder or a 
cylindrical gauze. By agitating the solution, raising the voltage and the cur- 

* Low, " Technical Methods of Analysis." 



COBALT 145 

rent density, the rate of deposition may be increased. In a properly agitated 
solution the deposition may be completed in forty-five minutes. 

The current should not be cut oflf until the solution is tested to determine 
if the electrolysis is complete. This is done by mixing a drop or two of the 
solution from the end of a stirring rod with a few drops of ammonium sulphide. 
If the electrolysis is complete the mixture will remain colorless, but if some 
cobalt still remains in the solution the mixture will be darkened. After the 
electrolysis is complete the cathode is carefully removed from the solution and 
dipped into a beaker of clean water, after which it is washed with alcohol, pref- 
erably ethyl alcohol. 

If a large number of electrolytic determinations are to be made, it is con- 
venient to have a wide-mouthed bottle with a well-ground-in glass stopper or a 
cork stopper for holding the alcohol for the preliminary washing. The mouth 
should be large enough to receive the cathode without pouring out the alcohol. 
The cathode may be lowered into the alcohol in this bottle, which should only 
be partly filled, and then rinsed again by pouring fresh alcohol over it and allow- 
ing it to drain into the wide-mouth bottle. This allows a great many cathodes 
to be washed with a comparatively small quantity of alcohol. Directly after the 
final washing with alcohol the cathode is passed through the flame of a Bunsen 
burner and the alcohol ignited. After this is entirely burned off the cathode 
is placed in a desiccator to cool and when cool is weighed. The increase in weight 
of the cathode is the weight of cobalt in the sample if the solution had been free 
from nickel before electrolysis. If the nickel remained in the solution the increase 
in weight of the cathode represents the cobalt and nickel in the sample. If it 
is desired to determine the cobalt and nickel together the increase in weight of the 
cathode is divided by the weight of the sample and multiplied by 100 to obtain 
the percentage. If it is desired to obtain the percentage of cobalt separately, 
the plate is dissolved from the cathode in a few cc. of nitric acid and the nickel 
determined in the resulting solution by precipitation with dimethyl-glyoxime 
as described in the chapter on Nickel, after which the cobalt is found by 
difference. 

Cobalt in Cobalt Oxide ^ 

One gram of finely ground cobalt oxide is either fused with 10 grams of potas- 
sium bisulphate or heated with 20% sulphuric acid until dissolved. If the 
fusion method is used the melt is extracted with water and acidified with sulphuric 
acid. Arsenic and copper are precipitated by passing hydrogen sulphide through 
the warmed solution, which should be diluted to about 200 cc. for about one 
hour. These are removed by filtration and the cobalt determinated by one of the 
above methods. The following procedure is one of the most satisfactory: 

Procedure. If it is desired to determine the nickel separately, as is usually 
the case, this is first precipitated with dimethylglyoxime as described in the 
chapter on Nickel, after boiling the solution to expel hydrogen sulphide. It is 
then evaporated to fumes of sulphur trioxide and taken up with twice its volume 
of water. The free acid is neutralized with ammonium hydroxide and an excess 
of 50 cc. of strong ammonium hydroxide added. The solution is made up to 
250 cc. and electrolyzed as under Precipitation of Cobalt by Electrolysis. 

* R. W. Landrum, Proc. Am. Ceramic Soc, 12, 1910. 



146 COBALT 



Cobalt in Metallic Cobalt and Ferro-cobalt 

Cobalt is usually determined in metallic cobalt and f ciiro-cobalt by electrol- 
ysis, after separation of the elements precipitated by hydrogen sulphide in acid 
solution and elimination of iron, if present in large quantities. In case it is 
desired to estimate nickel separately it is precipitated by dimethylglyoxime as 
described in the chapter on Nickel, before electrolysis, taking the solution down 
to sulphur trioxide fumes, diluting with water and adding ammonium hydroxide 
in excess and electrolyzing. In case the solution is electrolyzed before separ 
rating the nickel the determination of this element may be made in the solution 
of the electrolytic deposit dissolved in acid, the cobalt then found by difference. 

Procedure. Dissolve 1 gram of well-mixed drillings in the least possible 
quantity of nitric acid and add 20 cc. of 1 : 1 sulphuric acid. Evaporate to 
fumes of sulphur trioxide and allow to fume strongly for ten minutes. This 
insures the complete elimination of nitrates, which would interfere subsequently 
with the electrolysis. Cool and dilute carefully with 20 cc. of water. Heat 
the solution to nearly boiling and pass in hydrogen sulphide for one hour to 
precipitate copper and arsenic. Filter and boil the solution to expel the last 
traces of hydrogen sulphide. Add 2 cc. of hydrogen peroxide to oxidize ferrous 
compounds to ferric state, and add ammonium hydroxide until slightly alkaline 
fo litmus paper and heat to boiling. Filter off the ferric hydroxide and wash 
with water containing a small quantity of ammonium hydroxide. Redissolve 
the precipitate in a little 1 : 1 sulphuric acid, adding a little hydrogen peroxide 
to keep the iron in the ferric state, and reprecipitate in the same manner as that 
described above. In presence of comparatively large 'amounts of iron the basic 
acetate separation of iron is necessary, as Fe(OH)i occludes cobalt and nickel. 
The filtrates from these precipitations are added to the main one. 

In determining the cobalt in metallic cobalt it is not necessary to filter off the 
iron precipitate, if this is small, as it has been found by W. L. Rigg, of Deloro, 
Ontario, that this precipitate does not interfere with the accuracy of the deter- 
mination. The iron content may be up to 5% without interfering seriously with 
the electrolysis. 

The solution is made ammoniacal with 50 cc. of strong ammonium hydroxide 
and electrolyzed as described above. 

Cobalt in Metallic Nickel 

The cobalt in metallic nickel may be determined by precipitation with potas- 
sium nitrite from a solution of the sample containing an excess of acetic acid. 
The precipitate is filtered off and dissolved in hot sulphuric acid solution, after 
which the solution is evaporated to fumes of sulphur trioxide and carefully 
diluted. The excess of acid is neutralized and made strongly ammoniacal with 
ammonium hydroxide. The solution is then electrolyzed as previously described. 

Procedure. Dissolve 5 grams of thoroughly mixed drillings in a minimum 
quantity of nitric acid. Evaporate to a syrup. Care must be exercised at this 
point to prevent evaporating too far and decomposing the nitrates. Dissolve 
in 60 cc. of water. Neutralize with a practically saturated solution of sodium 
carbonate. For this purpose a dropping bottle is very convenient. Render 
slightly acid with acetic acid and add an excess of 10 cc. of 1 : 1 acetic acid. Heat 



COBALT 147 

to almost boiling and add 10 cc. of a 50% solution of potassium nitrite to which 
has been added 10 cc. of glacial acetic acid per 100 cc. of solution. This solu- 
tion must also be nearly boiling and should be added while gently rotating the 
nickel solution. Wash down the sides of the beaker with a 1% solution of 
potassium nitrite containing 1 cc. glacial acetic acid per liter. Allow to stand 
for at least six hours and preferably overnight. Filter through a thick, 9-cm. 
filter paper without previous wetting. Considerable care is required to keep the 
precipitate well down in the apex of the filter paper cone, as it creeps very badly. 
Wash about ten times with the warm nitrite solution mentioned above. Lift gently 
from the funnel and open the filter paper into a beaker. Lay the paper against 
the side of the beaker with the outside against the glass. This leaves the paper 
adhering to the side of the beaker in a most convenient position for washing. 
Wash down as much of the precipitate as possible with about 10 cc. of 1 : 1 sul- 
phuric acid solution, heated to about 80° C. This should leave only a slight 
film of precipitate on the paper. Keep the solution at about 80° C. and wash 
the paper five times with the warm sulphuric acid solution, using about 10 cc. 
each time, gradually withdrawing the paper from the top of the beaker. Give 
a final wash with hot water and squeeze the last drops from the filter paper into 
the beaker. Evaporate and allow to fume strongly for ten minutes. Add water 
in a cooling trough until the volume is about doubled. Neutralize with ammo- 
nium hydroxide and add an excess of 50 cc. of strong ammonium hydroxide and 
electrolyze as described in Precipitation of Cobalt by Electrolysis. 

Cobalt in Ores and Enamels^ 

The determination of cobalt in ores and enamels is usually made by a slight 
variation of the above methods. The silica is separated in the usual manner 
by taking down to dryness with hydrochloric acid and the warmed solution is 
treated with hydrogen sulphide to precipitate sulphides insoluble in acid solution. 
Aluminum, chromium and iron are precipitated by adding ammonium hydrox- 
ide to the oxidized solution. In the enamel industry it has been the practice 
to follow R. W. Landrum's method, in which the cobalt, manganese and nickel 
are precipitated together as sulphides and filtered off. The manganese is dis- 
solved from this precipitate with cold hydrogen sulphide water acidified with 
one-fifth its volume of hydrochloric acid (sp.gr. 1.11). The residue of cobalt 
sulphide is burned in a porcelain crucible, dissolved in aqua regia and evap- 
orated with hydrochloric acid. The platinum and copper, if they are present, 
are thrown down by passing hydrogen sulphide through the solution. The 
filtrate is made ammoniacal and the cobalt is precipitated with hydrogen sul- 
phide. This is filtered off and washed with water containing a small quan- 
tity of ammonium sulphide. The precipitate is either ignited and weighed as 
oxide or reduced in hydrogen to metallic cobalt, taking care to cool it thor- 
oughly in an atmosphere of hydrogen before allowing it to come into contact 
with the atmosphere of the room, as finely divided cobalt is decidedly pyrophoric 
and oxidizes readily, particularly if reduced at a low temperature. 

Instead of igniting the sulphide precipitate it may be dissolved in hot 1 : 1 
sulphuric acid solution with the aid of a little nitric acid and treated as described 
under Precipitation of Cobalt by Electrolysis. 

» R. W. Landrum, Trans. Am. Cer. Soc., 12, 1910. 



148 COBALT 



Cobalt in Steel 

This determination is a modification of the nitroso-beta-naphthol method 
abeady described, as worked out in the laboratory of the Firth Stirling Steel 
Company, McKeesport, Pa. The procedure as described by Mr. Giles, Chief 
Chemist, is as follows. 

Two grams of the sample are weighed into a 5(K)-cc. Erlenmeyer flask and 
dissolved in 50 cc. of concentrated hydrochloric acid. When the sample is 
completely decomposed 10 cc. of concentrated nitric acid are added to oxidize 
the iron, tungsten, etc. The solution is evaporated to 10 cc; 50 cc. of water 
are added; the contents of the flask are then transferred to a 500-cc. volumetric 
flask and cooled to room temperature. A fresh solution of zinc oxide is added 
m slight excess, the contents of the flask diluted to the mark, well mixed, trans- 
ferred back to the original Erlenmeyer flask and allowed to settle. Filter 250 
cc. (equivalent to 1 gram of the sample) through a dry filter paper, transfer it 
to a 500-cc. flask, then add 6 cc. of concentrated hydrochloric acid. 

The solution, which should now be between 300 and 350 cc. in volume, is 
heated to boiling and 10 cc. of freshly prepared solution of nitroso-beta-naphthol 
(1 gram of salt to 10 cc. glacial acetic acid) are added for each 0.025 gram of 
cobalt present. Continue to heat for two minutes, remove from plate, shake 
well, and set aside until the bright red precipitate settles, which will only take 
a few minutes. Filter the hot solution and wash the flask out with hot 1 : 1 
hydrochloric acid and then wash the flask out with hot acid of the same strength. 
Wash the paper alternately with hot (1 : 1) hydrochloric acid and hot water 
until it has been washed five times with the acid, then wash ten times with hot 
water. The precipitate is transferred to a quartz or porcelain crucible, heated 
gently to expel the carbonaceous matter, then at a high temperature until ignition 
is complete. After cooling the crucible is weighed and the weight of the residue 
(CojOi) is multiphed by 0.734 to obtain the percentage of cobalt present. If 
desired the C01O4 may be reduced in hydrogen and weighed as metal. 



COPPER 

Wilfred W. Scott and W. G. Derby. 

Cu, af.irf. 63.57; 9p.gr. 8.89^^''; m.p. 1083 (in air 1065); 6.p. 2310; 

oxides CU2O a¥id CuO. 

DETECTION 

Copper is precipitated in an acid solution by HjS gas, along with the other 
members of the hydrogen sulphide group. The insolubility of its sulphide 
in sodium sulpliide is a means of separating copper from arsenic, antimony, and 
tin. The sulphide dissolves in nitric acid (separation from mercury) along with 
lead, bismuth, and cadmiiun. Lead is precipitated as PbS04 by sulphuric acid 
and bismuth as the hydroxide, Bi(OH)s, upon adding ammonium hydroxide. 
Copper passes into the filtrate, coloring this solution blue, 

Cu(0H)2 • 2NH4OH • (NH4)2SO.. 

Flame Test. Substances containing copper (sulphides oxidized by roasting), 
when moistened with hydrochloric acid and heated on a platinum wire in the 
flame, give a blue color in the reducing flame and a green tinge to the oxidizing 
flame. 

Wet Tests. Nitric acid dissolves the metal or the oxides (sulpliides should 
be roasted), forming a green or bluish-green solution. Anmionium hydroxide 
added to this solution y^i^\ precipitate a pale blue compound, which dissolves 
in excess with the formation of a blue solution. (Nickel also gives a blue color.) 

Hydrogen sulphide, H2S, passed into an acid solution containing copper, 
precipitates a brownish-black sulphide, CuS. (Distinction from nickel.) 

Copper is displaced from its solution by zinc, cadmium, tin, aluminum, lead, 
bismuth, iron, cobalt, nickel, magnesium, and phosphorus. From a potassium 
hydroxide solution it is precipitated by KzSnOj. If a strip of iron is placed in a 
solution of copper, neutral or slightly acid, it will be coated over with metallic 
copper. (Delicacy 1 part Cu per 120,000 of solution.) 

The greenLsh-blue cupric salts in acid solution are reduced to the colorless 
cuprous compounds by metallic copper and by stannous chloride and by 
arsenious acid, grape sugar, sulphurous acid in alkaline solutions. 

ESTIMATION 

The estimation of copper is required in the following substances: In ores^ 
of copper f in which it occurs as native copper or combined as sulphide, oxide, 
carbonate, chloride, and silicate. In furnace slags, mattes, concentrates, blister 
copper, bottoms. The determination of copper is required in the analysis of 

*Ores, copper pyrites, yellow copper ore, CuFeS2; copper glance, CU2S (gray to 
bluish-black); malachite, CuCOsCuOHHjO (green): azurite, 2CuC08CuOH20 
(blue); cuprite, red copper, CutO; malaconite, CuO (black); dioptase, CuOSiO-HjO 
(green vitreous). 

149 



150 COPPER 

alloys containing copper,^ brass, bronze, etc. It is occasionally looked for as 
an undesirable impurity in food products. It is determined in salts of copper, 
in insecticides, germicides, etc. 

Preparation and Solution of the Sample 

Hydrochloric and sulphuric acids are effective in dissolving metallic copper 
only in presence of an oxidizing agent; nitric acid is the most active solvent. The 
oxides of copper may be dissolved in hydrochloric or sulphuric acid, but nitric 
acid is commonly used. 

Ores. If the ore consists practically of a single mineral, the fineness of the 
sample need not exceed 80 mesh. If the ore is a mixture of minerals, lean and 
rich in copper, the laboratory sample should pass a 120-mesh sieve. 

MetalUc particles or masses are separated at some stage in the process of 
sampling and made into a separate sample. If the metalUc portion is a small 
percentage of the total sample and consists of particles, the copper value of 
which is known to vary by a few percent, no attempt is made to refine the sample 
of such, but a large portion, 10-100 grams, is taken for analysis and the copper 
determined in an aliquot part of the solution. If the metallic masses are a large 
percentage of the sample, large of size, or consisting of particles differing widely 
in copper content, a weighed amount of 1 to 50 lbs. is melted in a graphite crucible, 
with addition of suitable fluxes, such as powdered silica or lime, if necessary. 
Separate samples are made of the weighed products of the fusion and the copper 
content of the material before melting calculated from their analyses. The amount 
of the sample taken for analysis depends upon the richness of the ore; as a general 
rule 0.5 to 1 gram sample is taken of ores containing over 30% copper, 2 grams 
of 10 to 30% copper ores and 5 grams of ores containing less than 10% copper. 

Sulphide Ores. Copper Pyrites, Copper Glance, Iron Pyrites, etc. One 
to five grams of the finely ground ore is dissolved in a flask by adding 10 to 20 
cc. of dilute nitric acid (sp.gr. 1.2), warming gently for about fifteen minutes. 
The solution is evaporated to small volume and nitric acid expelled by either 
taking to drj'^neas, after adding hydrochloric acid, or to SO3 fumes, upon the 
addition of 5 to 10 cc. of dilute sulphuric acid, 1:1. In presence of lead the 
latter procedure is reconunended. 

The residue is taken up with 20 cc. of water acidulated with sulphuric acid 
(10%) diluted to alx)ut 150 cc. and the mixture brought to boiling. Lead sul- 
phate, if present, is filtered off together with silica, and copper passes into the 
solution. 

Copper may now be separated from other interfering elements by one of 
the procedures outlined under Separations, then determined gravimetrically 
or volumetrically. 

Notes. The sulphur that appears upon adding acid to the ore, with proper 
precautions, should \)e yellow. If it is dark and opaque, the solution has been over- 
neated, and sonic of the ore has been occluded. It is advisable in this case to remove 
the globule of sulphur and oxidize it separately with bromine and nitric acid, then 
boil out the bromine and add the solution to the rest of the sample. 

Sulphide ores may be treated according to the procedure recommended for iron 
pyrites in the chapter on Sulphur, the ore being decomposed with a mixture of bro- 
mine and carbon tetrachloride, 2 : 3, followed by nitric acid and then sulphuric acid. 

^ Alloys of zinc, tin and zinc, aluminum, silver, nickel, manganese, and gold. 



COPPER 151 

Matte. 0.5 to 1 gram of the fine sample is dissolved in nitric acid and 
evaporated with sulphuric acid as in case of ores. 

Oxidized Ores, Oxides, etc. The sample is dissolved in nitric acid and 
evaporated with dilute sulphuric acid to pastiness, and then heated to SOj 
fumes. Frequently a direct treatment with dilute sulphuric acid or with hydro- 
chloric acid may be employed. 

Treatment of Matte Slag. Only by quick quenching of the molten slag is 
decomposition of the sample by acids made possible, without preliminary treat- 
ment with hydrofluoric acid. As a rule lime slags are readily decomposed by 
mixed acids. Extremely acid, or iron slags, are apt to be refractory and are 
decomposed with most certainty by treatment with hydrofluoric acid followed by 
fusion with potassium bisulphate. 

The following scheme (White — Chemist Analyst, July, 1912) of attack, which 
also can be applied to sihcious ores, with skilful manipulation gives very satis- 
factory results: 

One gram of the 100 mesh fine slag is placed in a 250 cc. beaker of Jena glass, 
moistened with water, mixed with 3 cc. of sulphuric acid (sp.gr. 1.54), and 
then, while the particles of the slag are in suspension through rotary movement 
of the beaker, 15 cc. hydrochloric acid are added. The sihca is gelatinized in 

2 or 3 minutes by heating the beaker over a free flame. One cc. nitric acid followed 
by a few drops of hydrofluoric acid are added, and the heating continued in a 
hood until the material is nearly dry, and then to strong sulphuric acid fumes 
on a hot plate. When cool, 4 cc. of sulphuric acid (sp.gr. 1.54) are added. 

The remainder of the procedure depends upon the method that is to be fol- 
lowed in the determination of copper. If the electrolytic method is preferred, 

3 cc. of nitric acid are added; the mass heated until solution is effected, the liquid 
diluted to 175 cc. with cold, distilled water, and copper plated out in 20-35 
minutes, using a rotating anode and 2i amperes current. 

If the iodide method is to be followed, without addition of other acid than sul- 
phuric, the mass is again heated to fumes. When cooled, 25-30 cc. water and 
5 cc. hydrochloric acid are added and the liquid boiled until clear. After addition 
of 40 cc. saturated solution of sodium acetate, 4}% solution of sodium fluoride is 
added until the color of ferric acetate is discharged, and then an excess of 10 cc. 
When cold, titration is commenced, using a thiosulphate solution with a copper 
equivalent of 0.0005 g. per cc. 

The following quick method has been systematically and satisfactorily checked 
for a long period by a hydrofluoric acid-bisulphate fusion method, by which cop- 
per, precipitated as a sulphide, is ignited, the oxide dissolved in nitric acid and 
copper determined by electrolysis. 

Three grams of the 100 mesh fine sample are placed in an 800 cc. Jena beaker. 
The slag is spread over the bottom of the beaker, and while in motion 5 cc. 
of sulphuric acid are added rapidly to prevent the slag gathering into a mass. 
After addition of 40 cc. hydrochloric acid, the beaker is heated over a bare flame 
for about 3 minutes until the silica has gelatinized. To the hot solution nitric 
acid is added, drop by drop, until the liquid becomes dark brown. To the liquid, 
while in a state of agitation, 1-2 cc. hydrofluoric acid are added and the mix- 
ture boiled until the solution is complete. The liquid is diluted to 400 cc. and 
saturated with hydrogen sulphide and the precipitate filtered and washed as 
usual. The copper sulphide is ignited in a sihca crucible; the residue, if washing 
of the precipitate has been thorough, can be brushed into a 250 cc. beaker dan 



152 COPPER 

dissolved with a few cc. of nitric acid. After boiling gently to expel nitrogen 
gases, the free acid is neutralized with ammonia, and the solution then acidified 
with a slight excess of acetic acid. The cold solution is titrated by the iodide 
method, using a thiosulphate solution having a copper equivalent of about 0.0005 g. 
per 1 cc. 

Metals. A casting of a copper alloy and even of refined copper is not homo- 
geneous, and the zones of segregation of the constituents of the alloy (usually 
roughly parallel to the cooling surfaces) are the more sharply defined as the 
conditions which favor diffusion of the eutectic prevail, therefore, unless the cast- 
ing be quite thin and quickly cooled, a satisfactorily representative sample of it 
cannot be obtained from a single drill hole. A single casting may be sampled by 
complete cross-sectional cuts by a suitable saw or by a series of drill holes located 
in such a manner as to amount substantially to one or more cross-sectional cuts. 
Steel is usually present as a contaminant of the drill or saw shavings from refined 
copper and the tougher alloys and should be removed by a magnet. Crude 
copper, such as blister or black copper, is sampled by drilling one hole in each 
piece of a definite fraction of the total pieces of the average lot. The position 
of the hole in successive pieces is changed to conform with a pattern or 
" templet " which will cover a quarter, or half, or the complete top surface of 
the average piece, the " templet " is divided into squares, preferably about 1 inch 
on a side, and in the centre of each square the i-inch hole is drilled. The 
drillings are ground to pass a 20-mesh screen and the sample then withdrawn by 
means of a riffle sampler. 

Sampling by splashing from a molten stream and by slowly pouring the 
metal into water are methods frequently practiced. The size of the particles, 
the degree of homogeneity and the limit of accuracy of result required are factors 
which determine whether one or more grams of the sample should be taken for 
analysis. 

Iron Ores and Iron Ore Briquettes. A 5-gram sample of the finely 
divided material is fused in a large platinum dish with 40 grams of pure potas- 
sium bisulphate. If the ore is high in sulphur, it should be roasted by heating 
to redness in a silica or porcelain crucible before placing in the platinum dish 
and mixing with the bisulphate. 

The cooled fusion is broken up into small pieces and placed in an 800-cc. 
beaker with clock-glass cover. Three hundred cc. of hot water and 25 cc. of 
strong hydrochloric acid are added and the fusion is boiled until it passes into 
solution. If an appreciable residue remains, the solution is filtered, the residue 
fused with additional bisulphate, then dissolved in hot dilute acid and the 
filtrate added to the first solution. Silica and barium sulphate remain in the 
residue. 

The solution is now reduced and copper precipitated according to directions 
given under *' Separation of Copper by Precipitation in Metallic Form by a 
more Positive Element," aluminum powder being preferably used. 

The precipitated copper is filtered free from iron and other conmionly 
occurring impurities, then dissolved by pouring on the precipitated metal 30 
cc. of hot dilute nitric acid, 1:1, followed by 10 cc. of bromine water and then 
10 cc. of hot water. The filter paper is removed, ignited and the ash added to 
the copper solution. The whole solution is now evaporated to small volume 
and determined, preferably, by the *' Potassium Iodide" method as described 
under the volumetric procedures. 



COPPER 153 

Steel, Cast Iron, and Alloy Steels.^ From 3 to 5 grams of steel, depending 
upon the amount of copper present, are dissolved in a mixture of 60 cc. of 
water and 7 cc. of sulphuric acid (sp.gr. 1.84) in a 250-cc. beaker. After all 
action has ceased, a strip of sheet aluminum, IJ ins. square, bent so that it will 
stand upright in the beaker, is placed in the solution. 

After boiling the solution for twenty to twenty-five minutes, which is suf- 
ficient to precipitate all of the copper in the sample, the beaker is removed 
from the heat and the cover and the sides washed down with cold water. The 
liquid is decanted through an 11 -cm. filter, the precipitate washed three times 
with water, then placed with the filter in a 100-cc. beaker, and 8 cc. of con- 
centrated nitric acid and 15 cc. of water are poured over the aluminum and 
the solution heated to boiling. This hot solution is poured over the precipitate 
and filter in the 100-cc. beaker, and boiled until the paper becomes a fine pulp, 
only a few minutes being required. The solution is filtered, the residue washed 
several times with hot water and the filtrate and washings, not over 200 cc, 
are received in an electrolytic beaker, 2 cc. of concentrated sulphuric acid added 
and the solution electrolyzed, using a current of 2 amperes with an E.M.F. 
of 2 volts. With gauze cathodes and anodes the deposition is complete in an 
hour and a half. 

SEPARATIONS 

Isolation of copper in presence of large amounts of iron, or in substances 
containing nickel, cobalt, zinc, bismuth, cadmium, etc., may be accomplished 
by precipitation of the element as cuprous sulphocyanate according to the 
following procedure: 

Precipitation of Copper as Sulphocyanate. Nitric acid having been 
expelled from the sample, the solution, 50-100 cc, is nearly neutralized with 
sodium carbonate and the copper reduced by addition of sodium bisulphite 
or metabisulphite or by passing in SOi gas. The solution is gently warmed 
and potassium sulphocyanate reagent added (50 grams KCNS salt per liter), 
until no further precipitation takes place. The sulphocyanate solution may be 
prepared by addition of 50 grams of potassium bisulphite or metabisulphite 
to the above reagent. The preliminary reduction of copper is then unneces- 
sary, as reduction takes place with addition of the reagent. After settling, 
the precipitate is collected on a filter and washed free of acid, first washing 
with the precipitating reagent, then with anunonium acetate solution and 
finally with water. 

The precipitate may now be dissolved in nitric acid and evaporated to near 
dryness with sulphuric acid and copper determined by electrolysis or by potas- 
sium iodide procedure. 

The precipitated cuprous sulphocyanate may be weighed after dr3ring at 
100** C, the compound having been collected in a weighed Gooch crucible. The 
compound multiplied by 0.5226 gives the equivalent metallic copper. 

The precipitate may be dried and burned with sulphur and the residue weighed 
as cuprous sulphide, CutS. This multiplied by 0.7986 gives the equivalent weight 
of copper. 

Reaction. 2CuS04+2KCNS+SO,+2H,0 =2CuCNS+2H2S04+K2S04. 

1 W. B. Price, Jour. Ind. Eng. Chem., Vol. 6, No. 9, p. 170. 



154 COPPER 

Note. Cuprous sulphocyanate is insoluble in water and in dilute hydrochloric 
acid. With the exception of silver, selenium and tellurium, cop{)er is the only metal 
that is precipitated in hydrochloric acid solution by potassium sulphocyanate, hence 
it may oe separated from other elements that would mterfere in its determination by 
this method. 

Separation of Copper by Precipitation in Metallic Form by a More Posi- 
tive Element. Metallic aluminum or zinc is more commonly used in this 
procedure. A strip of pure aluminum or zinc, placed in the neutral or slightly 
acid solution, causes the complete deposition of copper. The copper is removed 
mechanically from the displacing metal and dissolved in nitric acid and then 
estimated, or the aluminum may be dissolved with the copper. 

A method of precipitation by means of powdered aluminmn is recom- 
mended especially for separation of copper from large amounts of iron, iron 
ores and iron ore briquettes. The solution of the bisulphate fusion of the iron 
ore is heated until bubbles appear over the bottom of the containing beaker. 
Aluminum powder is now added in small portions at a time, in sufficient quantity 
to reduce the iron, the solution becoming colorless. The solution is now heated 
until the aluminum completely dissolves. Metallic copper is precipitated. It 
is advisable to add 25 cc. of water saturated with HjS gas to precipitate traces 
of copper in solution. * The solution is filtered while hot through a filter (S. 
& S. No. 589), and washed six times, keeping the residue covered with water 
to prevent oxidation by air. The copper is now dissolved in hot dilute nitric 
acid, evaporated to small volume and determined by the procedure preferred. 
The potassium iodide method gives excellent results. 

Separation of Copper from Members of the Ammonium Sulphide and 
Subsequent Groups by Precipitation as Copper Sulphide in Acid Solution. 
The solution containing free hydrochloric or sulphuric acid is saturated with HxS 
gas,^ the precipitated copper sulphide (together with the members of the group), 
is filtered and washed, first with water containing H2S and finally with a little 
pure water. The residue is dissolved in nitric acid and the resulting solution 
examined for copper. 

Removal of Silver. This element is precipitated as the insoluble chloride, 
AgCl, by addition of hydrochloric acid, and may be removed by filtration, copper 
passing into the filtrate. 

Removal of Bismuth. Upon adding ammonium hydroxide to a solution 
containing copper and bismuth the latter is precipitated as Bi(OH)a and may 
be removed by filtration. Copper passes into the filtrate as the double ammo- 
nium salt. Ammonium carbonate or potassium cyanide may be used instead 
of anmioniiun hydroxide. 

Removal of Lead. Lead is precipitated by sulphuric acid as PbS04 and 
may be removed by filtration, copper passing into the filtrate. 

Removal of Mercury. The sulphide of mercury remains undissolved when 
the precipitated sulphides are treated with dilute nitric acid, copper sulphide 
dissolving readily. 

Removal of Arsenic, Antimony, and Tin. These elements may be re- 
moved by dissolving their sulphides with a mixture of sodium sulphide and 

sodiiun hydroxide. Copper sulphide remains insoluble. 

• 

* Copper may be precipitated as the sulphide by nearly neutralizing the free acid 
with sodium hydroxide, warming the solution and adding crystals of sodium thiosul- 
phate. Upon boiling black sulphide of copper is precipitated together with free 
sulphur. 



OOPPEK 



15S 



I 



In an alloy tin and antimony niay l>e jjreeipitated as oxides by evaporation 
of the Bolution of iLb nlloy wit.ii strong nitric ac-id, copper remains in readily 
Boluble fonii. 

Separation from Cadmium. The sulpliides in a solution of dilute sul- 
phuric acid, 1:4, are boiled and II,S gas passed in for twenty minutes, the 
solution being kept at boiling teinpemture. Cadmium sulphide dissolves while 
copper sulphide remains unaffected. Tlie Holution is filtered hot, the air above 
the filter IwJng displaced by CO, to prevent oxidation. Traces of cadmium are 
removed by repeating the operation. (Method by A. W, Hofmann.) 



GRAVIMETRIC DETERMINATIONS OF COPPER 

Deposition of Metallic Copper by Electrolysis 

The electrolytic method of determining copper is the most accurate of the 
gravimetric methods. Tlus deposition may conveniently be made from acid 
solutions containing free nitric or sulphuric acid) or from an ammooiacal solution. 



^hKir mrvjL-a,-t . v^---* 


_ ^ 


i 


^ ■, -, 


..jf,B. ,, ., 












' I 






\*^***v 


1 



PlO. 27. — Terminal Case Showing Battery of Electrodes for Electrolytic Dpiiosilion 
of Copiier. 

The end sought by this method is to plate out all, except a trace, of the 
Kipper in the form of an evenly distributed, firmly adherent, very finely crystal- 
line deinsit, which is free from a weighablc amount of impurity. 



156 COPPER 

In ores, mattes, aUoys (from which lead has been removed as the sulphate 
by taking the solution to fumes with sulphuric acid) deposition by electrolysis, 
from a solution containing free sulphuric acid, is convenient. On the other 
hand, deposition from a nitric acid solution is advantageous under conditions 
where this reagent has been used as a solvent and evaporation with sulphuric acid 
is unnecessary. This is the case in the analysis of certain alloys and the deter- 
mination of copper from which impurities have been largely removed. Deposi- 
tion from an ammoniacal solution is recommended when the copper salt con- 
tains chlorides and it is desired to avoid evaporation with sulphuric acid. A 
chloride in an acid solution gives rise to a spongy deposit of copper, and endangers 
a solvent action on the anode and deposition of platinum on the cathode. 

Conditions other than the presence of precipitable impurities, which affect the 
character of the deposit are — quantity and concentration of copper, size and 
shape of electrodes, current density, uniformity of distribution of current to the 
cathode, volume, temperature and rate of circulation of the electrolyte, and 
concentration of oxidizing agents such as nitric acid and ferric salts. Inas- 
much as the change of one condition limits or makes possible or necessary a modi- 
fication of others, a large number of practicable combinations of conditions are 
possible. For discussion of these conditions reference is made to articles by 
Blasdale and Cruess, Jour. Am. Chem. Soc. Oct. 1910, 1264; and by Richards 
and Bisbee, Jour. Am. Chem. Soc, May, 1904, 530. 

By the feature of rate of deposition, electrolytic methods may be classified 
as " slow " or " rapid." The slow methods, with 12 to 24 hour periods of elec- 
trolysis, are practiced when extreme accuracy is required, or when the distribution 
of laboratory labor and time allowed for completion of the assays permit 
their economical employment. The electrolyte is a solution of sulphate salts of 
the metals present, ammonium sulphate or nitrate, and a quantity of free nitric 
acid, which varies with the amount of copper and ferric salts present, and the 
current density employed. The oxidizing effect of nitric acid is intensified by the 
presence of ferric ions.^ Electrolysis is carried out at room temperature, at cur- 
rent densities varying from ND/100, 0.15 to 0.5 amperes; and deposition on plain, 
corrugated, slit or perforated platinum cylinders from 0.75 to 2 m. diameter 
having 50 to 200 cm. depositing surface. A perforated cylinder permits freedom 
of circulation between the two surfaces of the electrode, the most even distri- 
bution of current density, and produces the most uniform coating of the foil. 
On account of the effect on the character of the deposit by oxygen lodging 
in regions of the cathode where the current density and circulation is least, the 
anode should be of such a form that all the gas liberated will be in the zone of 
maximum circulation. To procure uniform behavior under given conditions the 
size and shape of the electrolytic beaker should be such as to present the smallest 
practicable volume of electrolyte between the outer surface of the cylinder and 
the inside of the beaker. An unclosed seam or rivetted joint in a negative 
electrode will hold tenaciously salts which require extreme care to remove. It is 
probable that such recesses retain traces of the electrolyte underneath the coat- 
ing of copper. 

Rapid methods have a tendency to procure high results, resolution and 
mechanical loss through misting having been prevented. Deposition Ls hastened 
by increasing the rate of circulation and the current density. Circulation is 

* Larison, Eng. and Min. Jour. 84, 442. Fairiie and Boone, Elect and Met. Ind. 
6, 68.) 



COPPER 



157 



promoted by the use of the gauze cathode,* by rotating either cathode,' or by 
placing the vessel, containing the solution and electrodes, in a field of electro- 
magnetic force.' Quick deposition of a quality satisfactory for some classes of 
work is brought about by increase of current density upon an electrolyte heated 
to 50** to 80° C. In all the quick methods, the progress of electrolysis should be 
watched, and the cathode removed as soon as completion of deposition is de- 
tected by the evolution of gas about its surface. The completion of action is 
ascertained with greater certainty by addition of water to the electrolyte and 
observing whether the newly exposed surface of the cathode remains bright. 
When the electrolyte is hot or has a high acid content, detachment of the cathode 
should be preceded by removal of the electrolyte and simultaneously washing 
the cathode without interruption of the current. A syphon may be employed, 
water being added as the liquid drains from the beaker until the acid is removed. 

RAPID METHODS 






i 



^ 



^ 






z^ 






iijij 



i99 






2%''-7c.m. 






^ S}^ "-/^ cm. 



% 



S;R 






^^SSS^ 



v*9 ?rr 




4=; 



Sofi Ste^i 



(0 



05 



N »0 



Rapid Deposition of Copper— Solenoid Method of Heath ^ 

The solenoid is made by winding 500 turns of No. 13 B and S gauge magnet 
wire upon a copper cylinder 2J in. in diameter, 3i in. high, ^ in. thickness of 
metal. The cylinder is 
brazed water ti^t at the 
bottom to a 5i in. disc of 
^ in. soft steel. In this 
disc is a 1-in. hole for the 
insertion of a rubber plug, 
through which glass tubes 
may be inserted for inlet 
and outlet of air or water 
to cool the electrolytic 
beaker. A steel disc of 
the same size as the bot- 
tom and with an opening 
to fit is brazed to the top 
of the cylinder. The sole- 
noid thus made is suitable 
for a 300 cc. lipless beaker 
4| m. high and 2\ in. 
dLuneter. The solenoid 
coil may be in series in the 
electrolytic line or excited 
separately. 

The negative electrode is of gauze 40 meshes per linear tach, with a depositing 
surface of 100 cm. and is slit to permit quick removal from the electrolyte. 

^Stoddard, Jour. Am. Chem. Soc, 1909, 385. Price and Humphreys, Jour. Soc« 
Chem. Ind., 1910, 307. 

« Eng. and Min. Jour., 89, 89, 1910. 

• PrsJy, Jour. Am. Chem. Soc, Nov., 1907, 1592. Heath, Jour. Ind. Eng. Chem., 
Feb., 1911, 74. 

* Heath, Jour. Ind. Eng. Chem., Feb., 1911, 7a 



JJJIJIII^Ff I 



Solenoid for Rotation 
of ihe Electrolyte 

Fig. 28. 



VI 



168 COPPER 

Procedure. Five grams of the thoroughly cleaned copper sample are dissolved 
in the covered electrolytic beaker on a steam plate with 40 cc. of stock acid solu- 
tion composed of 7 parts (1.42 sp.gr.) nitric acid, 10 parts sulphuric acid (1.84 
sp.gr.) and 25 parts by volume of water. The temperature during the solution 
is kept just below the boiling point, 50 cc. of the stock solution is used for copper 
containing 0.03 to 0.1 per cent of arsenic, 60 cc. for material containing 0.11 to 
0.5 per cent arsenic. The electrolyte is diluted to 120 cc. A current of 4.5 
amperes is used for the electrolysis and the same amount employed to excite the 
solenoid. During the deposition a double pair of watch glasses cover tightly 
the beaker until the color of the electrolyte fades out, when they are rinsed and 
removed. In about 30 minutes a test for completion of deposition is made 
by withdrawal of 1 cc. onto a porcelain tile and treating with a few drops of 
freshly prepared hydrogen sulphide water. This test will detect the presence of 
0.000005 g. copper or more remaining in the solution. The determination is 
complete in two and a half hours. 

Notes. The advantage of the solenoid over any mechanical device for the rota- 
tion of electrodes is due to the prevention of loss by spraying from the anode, as the 
beaker can be covered with a double pair of watch glasses. 

Results range from 0.003 to 0.01 per cent higher than the author's slow method of 
assay of refined copper, and is due to platinum from the anode, which is corroded by 
the mfluence of heat, nascent nitrous acid and high current. 

Deposition from Nitric Acid Solution. The solution should not contain 
over 2-3 cc. of free concentrated nitric acid. If more than this is present, 
the solution is evaporated to expel most of the acid, the remainder neutraUzed 
with ammonia and the requisite amount of nitric acid- added. The solution is 
diluted to 100 cc, warmed to 50° or 60° C. and electrolyzed with a current of 1 
ampere and 2-2.5 volts. Two hours are sufficient to deposit 0.3 gram copper. Since 
nitric acid acts vigorously on copper, it is necessary to wash out the acid from the 
beaker before breaking the current. (See method for copper in alloys, page 175.) 

Deposition from an Ammoniacal Solution. Ammonium hydroxide is 
added to the solution containing copper until the precipitate, first formed, dis- 
solves. Twenty to twenty-five cc. of ammonium hydroxide (sp.gr. 0.96) are 
required for 0.5 gram copper or 30-35 cc. for 1 gram. Three to four grams 
of anunonium nitrate are added and the solution electrolyzed with a current 
of ND/100=2 amperes. The electrodes are washed, without breaking the 
current, until the ammonia and nitrate are removed. 

Lead, bismuth, mercury, cadmium, zinc and nickel should be absent from 
the anunoniacal solution. Arsenic is not deposited. Unless a very pure platinum 
anode is used, platinum may contaminate the deposit appreciably. Jena or other 
brand of zinc borate resistance glass should not be used for the electrolytic beaker. 

SLOW METHODS 

Electrolytic Determination of Copper in Blister Copper 

The sample should be no coarser than 20 mesh. Because fine particles are 
comparatively poor in copper, extreme care must be taken in drawing the 
portion for analysis to preserve the ratio of the coarse to fine. Some analysts, 
to avoid sampling error, sieve the coarse from the 40 or 60 mesh fine and either 
make a separate analysis of each weighed product, or weigh into a single test 
the due proportion of each. Others draw a large portion^ by means of a riffle 



COPPER 



159 



volumetric liaKk, 



I 



(Fig. 294 nr similar sampling device and from its solutinn i. 
pipette an aliquot part pquivuleiit to one or more grams. 

By the snmll piirtior) tnelhod inwiluhlp matter must be removed by filtration. 
WLen the sample contains an insignificant quantity of insoluble matter, the prac- 
tice is to deposit the silver with the copper and 
makeacorrerlion for its presence in accordance 
with t.iic result n[ the BiK'cr assay of the sample. 

By the ii«^ portion method, insoluble 
inattcraiidsi]ver,aR silver chloride. Is removed 
from the electrolyte by sedimentation in the 
volumetric flask. 

Procedure. Small Portioii Method. The 
coarse and fine portions are quartered down 
to convenient amounts and from (hese a 5- 
gram composite weighed, which cont^tins the 
coaise and fine portions in ratio of Ihcir |>cr- 
«entage weights. The sample is plnc&l in a 
350-ce. tall-form beaker, without lip and with 
Baring riin. Fifty cc. of chlorine-free, slock 
Mcid solution {15 parts oilric and 5 piirt.« f<\il- 
phuric acids) are added, the beaker co\'ered 
with a funnel (stem up), which just fits in the 
rim, and the mixture heated gently at first and 
finally to l.H>iling. When the sample haa 
dissolved, 5 cc. saturated solution of am- 

nitrate are added and tlie sample diluted to 200 t 




Fio. 29.— Riffle Sampler. 

with water. 
When the electrolyte has cooled to room temperature the electrodes are intro- 
j duced , the beaker covered with s|ili t watch glasses and electrolysis started with a cur- 
' Tent of .05 ampere and continued until the appearance of the foil indicates that the 
silver has deposited. The current is then raised to ND/ 100 = .75 ampere and this 
continued far twenty to twenty-t»'o hours, or until the appearance of gas about the 
negative electrode indicates that deposition of the copper is practically complete. 
For the unexperienced a simple method is to add a little water to the electrolyte 
without breaking the current and after 15 minutes to observe whether any depo- 
sition or copper takes place on the freslily exposed surface. The watch glasses and 
electrode stems should be riiiaed when the electrolysis has continued 15-16 hours. 
Procedure. Large Portion Method.' The sample is quartered by a 
• riffle sampler (see Fig. 29) to an amount very close to >0 grams. This quan- 
! tity is weighed and transferred by a paper chute into a 2000 cc, flask, which 
has been calibrated by the method of repeated delivery at constant temperature, 
of a 50 cc. overflow, dividing pipette (see Fig, 30). The liquid employed in caL- 
brating is a copper solution of the same composition as that for which the 
flask is to be used. A cold mixture of 80 cc. sulphuric acid (spgr. 1.82) and 
200 cc. nitric acid (1,42) with 500 cc. of water is added. Astandard solution 
of sndiuiD chloride is added in sufficient quantity to precipitate the silver, care 
being taken to add less than 20% excess. A bulbed condenser tube is placed 
in the neck before putting the flask on a hot plate. 

The solution is gradually heatod to boiling and when the solution is nearly 
I complete, boiled gently for one hnur. This generally completely dissolves the 
■ W. C. Ferguson, Jour. Ind. and Eag. Cbem., May, 1910. 




COPPER 



copper present. Residuea of lead, tin, silver, or silica if present io appreciable 
amounts are separated at this point by filtration. 

When the solution in the flask has cooled for half an hour, water is added 
to a little above the 2000-cc. mark, giving the flask a rotary motion, during 
the addition, to mix the solution. The flask is placed in a large tank, Fig. 30, 




Fra. 30 — Conatant Temperature Bath and Dividing Pipette. 

containing water and aUowed to remain until it becomes of the same temperature 
as the water and 'very close to that of the room. The solution is then made 
exactly to the mark and allowed to settle, after thorough mixing, by placing 
the flask again in the water tank 

Electrolysis. Portions equivalent to 2 grama of san^ple are measured out 
by means of a dividing pipette, with wat«r-jacket through which 
the tank-wat«r flows. The solution is run into glasses, hydro- 
meter-jar in shape, with concave bottoms, height of glass, 6j ins., 
diameter 2j ina.. Fig. 31, Each portion ia treated with 6 cc. of 
a saturated solution of ammonium nitrate and diluted to 125 cc. 
with water. (NHiNO, or (NHi)jSO, delays deposition of Aa and 
Sb until electrolyte is freed from Cu.) The electrolyte, at this 
stage, contains about 3.7 cc. of nitric acid. 

The copper is deposited by electrolj'sis, using a current of 
.33 ampere per 100 sq.cm., which is kept constant until deposi- 
tion is complete, about twenty hours. It is advisable to begin 
the electrolysis in the evening, 5 p.m. The following morning, 
the inside of the jar, the rods of the electrodes, and the split 
watch-glasBes which cover the jar are rinsed with a spray of water into the 
glass and the run continued for two or three hours. Each electrode is quickly 




Fig. 31. 



COPPER 161 

detached from the binding posts, the cathode plunged into cold water, then 
successively into three jars of 95% alcohol, shaken free of adherent drops and 
dried by revolving rapidly over a Bunsen flame for a few seconds after ignition 
of the £Um of alcohol. 

The weighing of foil plus the deposit is made with as little delay as possible. 

Determination of the Copper Remaining in the Electroljrtes. The elec- 
trolyte is concentrated and any residual copper precipitated as sulphide by 
HsS after first neutralizing the free acid and then making sUghtly acid with 
HCl. The copper sulphide is dissolved with a little hot HNOj and made ammo- 
niacal. The color of the solution is compared with a standard solution treated 
with the same amount of reagents as the sample, care being taken that similar 
conditions prevail when making comparison. The electrolytes seldom contain 
over 0.01% copper. 

Notes and Precautions 

Character of the Deposits. The ideal deposit is of a salmon-pink color, silky 
in texture and luster, smooth and tightly adherent. A slightly spongy and coarsely 
crystalline deposit, although good in color and perfectly adherent, will invariably 
give high results. A loosely adherent deposit caused by either too rapid a deposition 
at the commencement or too low a current density at some period of the electrolysis, 
usually shows a red tint and may give a high result on account of oxidation or a low 
result because of detachment of particles. A darkly shaded deposit indicates the pres- 
ence of impurity in greater or less extent. If it is impossible to complete the electrol- 
ysis without this appearance the electrolyte should be purified. Impurities such as 
arsenic, antimony, bismuth, selenium and tellurium may occur in the blister copper. 

A dark colored, but perfectly adherent deposit is dissolved very slowly from the 
foil, in a covered electrolytic jar, by gently heating for several hours with about 60-70 
cc. of a solution containing 2 cc. sulphuric and 5 cc. nitric acids. When the solution 
is comi)lete the temperature is raised to expel dissolved a^sea. Five cc. saturated 
ammonium nitrate solution is added and the electrolyte diluted to 125 cc. When 
cooled to room temperature, electrolysis is carried out under the same conditions as 
th^t of the first deposit ana on the same foil, if arsenic or antimony is the interfering 
impurity; on a fresh foil if selenium or tellurium has been the contaminating element. 
The imdeposited copper is determined colorimetrically in the mixture of the first and 
final electrolytes and added to the weight of the copper deposited. 

If the sample contains a large percentage of arsenic or antimony, a portion represent- 
ing 2 grams is drawn from a pipette into a Kjeldahl flask, 10 cc. of sulphuric acia added, 
and the hquid boiled until nitric acid has been expelled. From this solution cuprous 
sulphocyanate is precipitated according to the method described on page 153. The 
funnel containing the filter is placed in a 500 cc. flask with long neck, the filter is punc- 
tured and the precipitate washed into the flask with the least quantity of water possible, 
the adherent precipitate is dissolved from the filter with warm dilute nitric acid, added 
cautiously to avoid violent evolution of gases from the dissolving precipitate in the 
flask. Tne washed filter is incinerated andthe solution of its ash by mtric acid reserved 
for addition to the electrolyte after completion of electrolysis. When solution of the 
precipitate is complete, the Uquid is boiled to small volume, neutralized, and 5 cc. 
anmionium nitrate solution and 3 cc. excess free nitric acia added. The liquid is 
transfered to an electrolytic jar and electrolysis carried out in the manner already 
described. 

The amounts of bismuth, arsenic, antimony, selenium or tellurium usually found 
in bUster copper may be precipitated together with iron present by addition of ammonia 
to a pipetted portion. The filtered precipitate is purified of copper by solution with 
nitric acid and reprecipitation. The combined filtrates are neutraUzed, 3 J cc. of free 
nitric acid added and the solution electrolyzed under the conditions already described. 
"Hie nitric acid solution of the incinerated filter, carrying the iron, etc., is added to the 
electrolyte after electrolysis is complete. Tne undeposited copper is determined 
colorimetrically by one of the procedures outlined on pages 165, 166 or 167. 

The deposited copper is never absolutely pure. The total impurities seldom 
exceed 0.03%. Ag from 0.000 to 0.18%; As from 0.000 to 0.003%; Sb from 0.000 



162 COPPER 

to 0.004%; Se and Te from 0.001 to 0.027%; Bi from 0.000 to 0.0003%. Periodical 
complete analyses may be made and corrections applied to the analysis when ex- 
ceedingly accurate percentages are required. 

Too low a current density or excessive oxidizing power of the electrolyte may pro- 
duce high results, due to the oxidation of the deposited copper. Too high a current 
density or a deficiency of oxidizing power in the electrolyte, by causing a deposition 
of impurities, will give hish results. 

The electrodes used by the Nichols Copper Co. are straight platinum wires for 
the positive ends and cylinders If in. long, 1 in. in diameter of 0.004 in. irido- 
platinum foil, 11} sq. in. depositing surface, lor the cathodes. 

A uniform current is essential. 

The nitric acid used should be free of iodic acid. 

The presence of oxide of nitrogen gases, or a chloride in an acid solution, will cause 
a coarsely crystalline or brittle deposit, under conditions which in their absence would 
produce a good plating. The deposit moreover may contain platinum from the anode 
if the electrolyte contains a chloride salt. 

Silver may be removed from the electrolyte by filtration^ upon precipitation as a 
chloride, or it may be deposited with the copper and correction made for its presence 
from the result of a separate assay. In the latter case the copper deposits in poor 
form, unless the silver be first plated out at a very low current density. 

Solid matter, unless removed, will contaminate the deposit mechanically. 

Arsenic, antimony, selenium or tellurium have an influence on the physical character 
of the deposit which may affect the copper result beyond the sum of such impurities 
depositee!. 

In the process of preparing an electrolyte, arsenic may be eliminated as arsenious 
fluoride in the decomposition of silicious material by hydrofluoric acid. Selenium is 
expelled by evaporation to dryness of a hydrochloric acid solution or by fuming a sul- 
phuric acid solution. All impurities may he removed by occlusion with ferric hydroxide; 
several times their weight of iron being added and the nydroxide then precipita,ted with 
ammonia. In the handUng of copper solutions account is to be taken of the retention 
of copper in the ferric hydroxide precipitate and the combination of copper in ammo- 
niacal solution with cellulose. 

Whether impurities are deposited or not, appreciably high results are obtained by 
continuing electrolysis for some time after the electrolyte has become impoverished 
of copper. 

Overheating of the copper deposit, in the process of ignition of the alcohol clinging 
to the cathode, will cause oxidation of the copper. As much as possible of the alcohol 
must be shaken off before passing the electrode rapidly through the flame. It is advis- 
able to weigh the copper shortly after deposition, as prolonged contact with air is unde- 
sirable, if extreme accuracy is desired. 

The copper deposits may l)e removed by plunging the electrode, for a few moments, 
in hot nitric acid. After washing with w^ater, the foil is ignited to a cherry red in a 
direct colorless flame. The ignition removes any grease which would \>e objectionable, 
that may contaminate the platinum. Alcohol frequently contains oily matter which 
will ding to the electrode in spite of the rapid ignition for drying the deposit. 



OTHER METHODS 

Determination as Cuprous Sulphocyanate 

The procedure has been outlined under Separations on page 153. 

CuCNS X 0.5226 =Cu. 

Determination as Copper Oxide ^ 

The solution, free from ammonium salts and organic matter, is heated to 
boiling in a porcelain dish and pure potassium hydroxide solution added, drop 

*" Analytical Chemistry," Treadwell and Hall. 



COPPER 163 

by drop, until a permanent precipitate, dark brown in color, is formed. The 
solution is alkaline to litmus-paper. The precipitate is washed by decantation, 
transferred to the filter and washed with hot water free of alkali. The precipi- 
tate and filter are ignited in a porcelain dish, first gently and finally with the 
full heat of a Bunsen burner. The residue ia weighed as CuO. 

CuO X 0.7989 =Cu. 



VOLUMETRIC METHODS FOR THE DETERMINATION 

OF COPPER 

Potassium Iodide Method 

The procedure depends upon the fact that cupric salts when heated with 
potassium iodide liberate iodine, the cuprous iodide formed being insoluble in 
dilute acetic acid is thus removed, no reversible reaction taking place. 

ReactioFU. 2CuS04+4KI=2CuH-2K2S04+I,. 

The liberated iodine is titrated with standard thiosulphate. 

2Na,S20,+2I =Na2S40e+2NaL 

This method is exceedingly accurate. Very few metals interfere. Bismuth, 
selenium, trivalent arsenic, antimony or iron should not be present. Lead, mer- 
cury, and silver increase the consumption of iodide, but do not otherwise interfere. 

Solutions. Sodium Thiosulphate. 7.5 grams of the salt, Na^SsOs -51120, 
are dissolved and made to 2 liters with water. The solution is standardized against 
a copper solution containing 1 gram of pure copper per liter, 1 cc. =0.001 gram 
Cu. Approximately the same amount of copper is taken as will be determined 
in the ores. For high-grade copper ores and crude copper, etc., it is advisable 
to prepare a standard thiosulphate solution ten times the above strength. The 
copper solution is made slightly ammoniacal and then acid with acetic acid. 
Potassium or sodium iodide crystals, free from iodate, are added and the liberated 
iodine titrated with the standard thiosulphate. (See Procedure.) 

p-. . , : — : = value of 1 cc. of the thiosulphate solution. 

cc. thiosulphate required 

Standard Copper Solution. One gram of purest electrolytic copper is dis- 
solved in 20 cc. of dilute nitric acid, sp.gr. 1.2, and the solution diluted to 1000 
cc. For standardizing the thiosulphate to be used with high-grade copper 
ores, crude copper, blister copper, etc., a copper solution containing ten times 
the above amount of metallic coppers is prepared. 

The following additional reagents are requh^d: starch solution, solid potas- 
sium iodide, 50% acetic acid solution, and other common laboratory reagents. 

Note. Sodium thiosulphate is apt to change in strength upon standing, so 
that restandardization is nccessar>'- 

Procedure. The solution containing the copper, separated from inter- 
fering elements, by precipitation with aluminum powder or potassium sulpho- 
cyanate, is evaporated to about 30 cc. and the free acid neutralized with sodium 
carbonate, or ammonia, and then made slightly acid with acetic acid, 1 : 3, 
the solution becoming clear, about 3 grams of potassium iodide, or the equivalent 



164 COPPER 

of a saturated solution, are added and the liberated iodine titrated with standard 
thiosulphate, the reagent being added until the brown color changes to light 
yellow and after the addition of starch solution until the blue color fades out. 
The end-point is very sharp. 

Cc. thiosulphate multiplied by value of reagent gives weight of copper in 
sample. 

Notes. Nitrous oxides should be expelled before neutraUzing with alkalies. A 
large excess of acetic acid should be avoided. The solution should be cool and con- 
tain at least 6 parts of KI for 1 of Cu, e.g., 1 Cu = 5.2231 KI = 1.9965 1 = 3.9034 
Na^jOj -51110. The solution should be concentrated, 40 to fijO cc. 

Prof. Gooch recommends a volume of 100 cc, containing no more than 3 cc. 
nitric, sulphuric, or hydrochloric acids, or 25 cc. of 50% acetic acid, with 5 grams 
potassiimi iodide. Two to 3 grams more of potassium iodide are adaed if the titra- 
tions are large. "Methods in Chemical Analysis." 

When ferric iron is the only disturbing impurity and no nitrates are present, the 
necessity of separation of copper may be avoided by fixing the free mineral acid by 
use of sodium acetate and then adding a clear, 4^ per cent solution of sodium fluoride 
imtil the red color of ferric acetate has bleached and then an excess of 10 cc. (Jour. 
Sci. Chem. Ind., May 15, 1915, p. 452; Mott, Chemist Analyst, July, 1912.) 

Arsenic or antimony when present in trivalent form may be oxidized by treatment 
with bromine, chlorine, hydrogen peroxide or potassium permanganate, care being 
taken to expel or reduce any excess of the oxidizmg agent before titration. 

Potassium Cyanide Method 

This procedure is largely employed on account of its simplicity, although 
it does not possess the degree of accuracy of the Iodide Method. The procedure 
depends upon the decoloration of an ammoniacal copper solution by potassium 
cyanide. 

The operations of the standardization of potassium cyanide and of making 
the assay should be as near alike as possible. If iron is present in the assay 
it should be added to the standard copper solution titrated, in order to become 
accustomed to the end-point in its presence. 

Silver, nickel, cobalt, cadmium, and zinc interfere and should be removed 
if present in appreciable quantities. Precipitation of metallic copper by alumi- 
num powder, as directed under Separations, is recommended as a procedure 
for iron ores and briquettes. In presence of smaller amounts of iron, the titra- 
tion may be made in presence of iron suspended in the solution. It is not 
advisable to filter off this precipitate, as it invariably occludes copper. With 
practice, the shade of color the iron precipitate assumes at the end of the reac- 
tion serves as an indicator, so that the operator is assisted rather than retarded 
by its presence.* 

2Cu(NH,)4S04-H,0+7KCN = 

K,NH4Cu2(CN),+NH4CNO+2K2S04+6NH,+H,0. 

Standard Potassium Cyanide Solution. Thirty-five grams of the salt are 
dissolved in water, then diluted to 1000 cc. 

Standardization. 0.5 gram of pure copper is dissolved in a flask by warming 
with 10 cc. of dilute nitric acid (sp.gr. 1.2), the nitrous fumes expelled by boiling, 
the solution neutralized, diluted and titrated as directed under Ptocedure. 

» Sutton, " Volumetric Analysis." Davies, C. N.. 68, 131. J. J. and C. Beringer, 
C. N., 49, 3. Dr. Steinbeck, Z. a. C, 8, 1; C. N., 19, 181. 



COPPER 165 

If iron is present in the samples titrated, it is advisable to add iron to the 
standard copper solution as directed above. 

0.5 

=wt. Cu per cc. of standard KCN. 



cc. KCN solution 



Procedure. The solution containing the copper is neutralized with sodium 
carbonate or hydroxide, the reagent being added until a slight precipitate forms. 
One cc. of ammonium hydroxide is now added and the solution titrated with 
standard potassium cyanide solution. The blue color changes to a pale pink; 
finally a colorless solution is obtained. In presence of iron, when the copper 
is in excess of the cyanide, the iron precipitate possesses a purplish-brown color, 
but, as this excess lessens, the color becomes lighter until it is finally an orange 
brown, the solution appearing nearly colorless. The reagent should be added 
from a burette drop by drop as the end-point is approached. 

Cc. KCN X factor per cc. = weight Cu in assay. 



COLORIMETRIC DETERMINATION OF SMALL AMOUNTS 

OF COPPER 

Potassium Ethyl Xanthate Method 

The method is based upon the fact that potassium ethyl xanthate produces 
a yellow-colored compound with copper. The reagent added to a solution 
containing traces of copper will produce- a yellow color var>''ing in intensity in 
direct proportion to the amount of copper present. Larger amounts of copper 
with the reagent produce a bright yellow precipitate of copper xanthate. Small 
quantities of iron, lead, nickel, cobalt, zinc, or manganese do not interfere. The 
procedure is especially valuable for determination of the purity of salts crys- 
tallized in copper pans. 

Special Solutions. Stock Solution of Copper Sulphate. 3.9283 grams 
CuSOi-SHjO are dissolved in water and made up to a volume of 1000 cc. One 
cc. is equivalent to 0.001 gram Cu. 

Standard Copper Sulphate. Ten cc. of the stock solution are diluted to 1000 
cc. with distilled water. One cc. =0.00001 gram Cu. 

PotaBsium Ethyl Xanthate Solution. One gram of the salt is dissolved in 
1000 cc. of water. The solution is kept in an amber-colored glass-stoppered 
bottle. 

Procedure. Five grams of the substance are dissolved in 90 cc. of water 
(see note) and the solution poured into 100-cc. Nessler tube; 10 cc. of the potas- 
sium xanthate reagent are added and the solution mixed by means of a glass 
plunger. To a similar tube containing 50 or 60 cc. of water are added 10 cc. 
of the xanthate reagent and then gradually drop by drop the standard copper 
solution from a 10-cc. burette (graduated in -^ cc.) until the colors in both 
tubes match. 

If a = grams of the substance taken for analysis, 6=nimiber of cc. standard 
copper solution required to match the sample; then 6X0.00001 X lOO-^-a =% Cu. 

Notes. The amount of the substance to be taken varies according to its copper 
content. The greater the copper contamination of the salt, the less sample required. 
The solution should be neutral or only very slightly acid. 



166 COPPER 

In place of the Nessler tubes the special colorimetric apparatus described under 
Titanium and under Lead may be used. A veiy weak copper standard will be 
required for the comparison tube. 

If the substance is insoluble in water the copper is rendered soluble by treat- 
ment with nitric acid. Hydrochloric acid is added and the nitric expelled by evapo- 
ration. The substance is taken up with water and the insoluble resioue filtered off. 

Starch and organic matter are destroyed by addition of 10 cc. 10% sodium 
hydroxide +10 cc. of saturated sodium nitrate solution, then evaporating to dryness 
and igniting. Hydrochloric acid is now added to expel the nitric acid as direct^ 
above. 

Ferrocyanide Method for Determination of Small Amounts 

of Copper 

By this colorimetric method it is possible to detect one part of copper in 
2,500,000 parts of water. The procedure depends upon the purplish to chocolate- 
brown color produced by potassium ferrocyanide and copper in dilute solutions. 
The procedure is applicable to the determination of copper in water and may 
be used in presence of a number of elements that occur in slags. Iron also 
produces a colored compound with ferrocyanide (1 part Fe detected in 13 million 
parts HjO), so this element must be removed from the solution before testing 
for copper. 

Solutions. Standard Copper Solution. 0.393 gram CuSOi-dHjO per liter. 
1 cc. =0.0001 gram Cu. 

Ammonium Nitrate. 100 grams of the salt per liter. 

Potassium Ferrocyanide. Four grams of the salt per 100 cc. of solution. 

Procedure. A volume of 5 to 20 drops of potassium ferrocyanide, accord- 
ing to the amount of copper present in the solution, is placed in a tall, clear, 
glass cylinder or Nessler tube of 150 cc. capacity, 5 cc. of ammonium nitrate 
solution added and then the whole or an aliquot portion of the neutral * solution 
of the assay. The mixture is diluted to 150 cc. The same amount of ferrocyanide 
and ammonium nitrate solutions are poured into the comparison cylinder, placed 
side by side with the one containing the sample, on a white tile or sheet of white 
paper. The standard copper solution is now run from a burette into the 
comparison cylinder, stirring during the addition, until the color matches that 
of the assay. The number of cc. required multiplied by 0.0001 gives the weight 
of copper in the sample contained in the adjacent cylinder. 

Amount of Cu XI 00 ^w ^ . , , 

■r^T: — ; \ : = % Cu m the sample. 

W t. of sample compared 

Notes. The solution must be neutral, as the copper compound is soluble in anuno- 
nium hydroxide and is decomposed by the fixed alkalies. If the solution contains free 
alkalies, it is made slightly acid and then the acid neutralized with ammonia, added 
in slight excess. This is boiled to expel the excess of ammonia, and then tested accord- 
ing to the directions under " Procedure." Solutions containing free acids are neu- 
tralized with ammonia. 

Iron may be removed by precipitation with ammonia. As this hydroxide occludes 
copper, the precipitate should be dissolved and reprecipitated to recover the occluded 
copper. 

Determination of copper in water is accomplished by evaporating a quantity 
of water to dr>'nes8, takmg up the residue with a little water containing 1 cc. nitric 
acid, the resicfue having been ignited to destroy organic matter, precipitating iron 
with ammonia, as directed above, and determining copi)er in the filtrate. 

The colorimeter used in determination of traces of lead and for the colorimetric 
determination of titanium may be employed in place of the Nessler tubes. 



COPPER 167 

Ammonia Method for Determining Small Amounts of Copper 

In the absence of organic matter, nickel and elements giving a precipitate 
with ammonia, copper to an upper limit of 10 milligrams can be determined by 
comparison of the depth of the blue tint of its ammonium solution with a tem- 
porary or permanent standard copper solution of equal volume. Permanent 
standard solutipn of copper sulphate, free of nitrate, if kept cool and away from 
the direct sunlight, lasts for a long time.^ 

Hydrogen Sulphide Method 

In the absence of elements precipitated by hydrogen sulphide, copper to the 

limit of about 1 milligram, in a solution not too strongly acid with sulphuric or 

hydrochloric acid, may be determined by comparison of its sulphide with that of 

a known quantity of copper in equal volume and similarly treated. The liquid 

should be cold and the passage of the hydrogen sulphide stopped before the 

compound coagulates. 

Note. Either the ammonia or the hydrogen sulphide method is applicable to the 
determination of the copper not deposited in the operation of the electrolytic method. 

DETERMINATION OF IMPURITIES IN BLISTER AND 

REFINED COPPER 

Introduction. In the complete analysis of copper the following impurities 
are generally estimated: silver, gold, lead, bismuth, arsenic, antimony, selenium, 
telluriiun, iron, zinc, cobalt, nickel, oxygen, sulphur, and less commonly, tin 
and phosphorus. In high grades of blister and in refined copper the percentage 
of these impurities is very low, the blister copper usually averaging over 99.0% 
copper with silver and the refined copper over 99.93% of the metal. The principal 
impurity in the refined element is oxygen, which may be present to the extent 
of .02 to .15%, the remaining impurities being in the third decimal place. 
From this it is readily seen that large samples are required for the accurate 
determination of these constituents. The amount of sample taken in blister 
copper depends upon the grade of copper analyzed. The impurities in this 
vary from tenths of a per cent to thousandths, as the metal from one locality 
may contain quite appreciable amounts of a constituent, which may be present 
only in extremely small quantities or not at all in copper from a different section. 
In usual practice it is customary to take from 10 to 50 grams of blister and 50 
to 600 grams of refined copper for analysis, depending upon the purity of the 
material. If a larger sample than 60 grams is taken, it is necessary to divide 
the material into several lots, and, after removal of the bulk of copper and isola- 
tion of the impurities, to combine the filtrates or residues containing the con- 
stituents sought. 

In the procedures the smallest amount of sample, 10 grams, is taken as 
the basis of calculation for amounts of reagents used. For larger samples, in 
the initial treatment for removal of copper, proportionately larger amounts of 
the reagents are required, i.e., multiples of from 2 to 6 times the amount stated. 
A 60-gram sample is the largest amount of material handled in one lot. 

Scrupulous care must be exercised throughout the analysis to prevent con- 

^ Heath, Jour. Am. Chem. Soc, 19, 21. 



168 COPPER 

lamination of the sample or reagents, and to avoid loss of constituents. The 
reagents used should be free from the substance sought or from interfering sub- 
stances. It is the practice to carry blank tests of the reagents through under 
conditions similar to a regular analysis for iron, lead, zinc, arsenic antimony and 
sulphur. 

It is found best to determine the impurities in several portions, i.e., gold 
and silver by assay; bismuth and iron in one portion; lead, zinc, cobalt, and 
nickel in a second; arsenic, antimony, selenium, and tellurium in a third; and 
separate portions for sulphur, oxygen, phosphorus and tin, when these are 
occasionally required. 

Determination of Bismuth and Iron 

Separation of Copper. Amount of Sample. Blister copper 10 to 25 grams, 
refined copper 100 to 500 grams. The drillings are dissolved in a large beaker 
in 40 cc. of nitric acid per 10-gram sample and the free acid expelled by boiling. 
The solution should not become basic during the evaporation. Water is added 
to make the volume 130 cc. per 10 grams or proportionately more for larger 
samples. Ammonia is now added in sufficient excess to hold the copper in 
solution and 5 cc. of saturated ammonium carbonate solution and the sample 
diluted to 200 cc. (25 cc. (NHOsCOg per 50 grams, and dilution to 1000 cc). 
The beaker is placed on the steam bath for several hours, preferably over night. 
The solution is filtered hot (to avoid crystaUization of the copper salt), the first 
100 cc. being refiltered, and the residue washed with hot water containing a 
little ammonia. By this procedure the copper passes into the filtrate and 
bismuth and iron remain in the residue on the filter. 

Separation of Iron and Bismuth. The precipitate is dissolved in warm, 
dilute hydrochloric acid (1 : 3), ammonia added to the solution in sufficient 
amount to almost neutralize the acid and the solution then saturated with hydro- 
gen sulphide. After settling some time, the precipitate containing bismuth sul- 
phide is filtered off, iron passing into the solution. 

Determination of Iron. Hydrogen sulphide is ex-pelled by boiling the 
filtrate, and iron oxidized by addition of hydrogen peroxide, or potassium 
chlorate (nitric acid should not be used). The solution is evaporated to dry- 
ness and iron then determined in the residue by the stannous chloride method, 
details of which may be found in the chapter on Iron, page 221. 

Determination of Bismuth. The sulphides remaining on the filter are 
dissolved in nitric acid, the solution evaporated with sulphuric acid to SOs fumes 
to expel pitric acid, the concentrate diluted with water, and lead filtered off. Bis- 
muth is precipitated in the filtrate by addition of ammonia in slight excess, 
followed by 10 cc. of a saturated solution of ammonium carbonate, and boiling. 
The precipitate is settled for several hours or over night if preferred, and then 
separated by filtration. This is now disvsolved in the least amount of nitric 
acid, added to the filter drop by drop from a burette and bismuth determined 
in the solution by the cinchonine iodide method, givon in detail in the chapter 
on Bismuth, page 69. 

Notes. An excess of nitric acid, or the presence of cadmium, lead, silver, or 
hydrochloric acid interferes with the colormietric procedure. 

In analysis of refined copper several 50-gram portions are taken for analysis, 
ten such portions on a 500-gram sample; the filtrates, obtained upon dissolving the 
residue freed from copper, are combined and bismuth and iron aetermined on this 
combined solution. 



COPPER 169 



Determination of Lead, Zinc, Nickel, and Cobalt 

Removal of Copper. Ten to 25 grams of blister copper, and 100 to 250 
grains of refined copper in 25-gram portions are taken for analysis. The metal 
is dissolved in nitric acid (40 cc. per 10 grams) and the solution boiled until 
a faint green precipitate begins to appear on the surface of the solution. 
The free acid being expelled, the solution is made faintly acid by adding 1 to 
2 cc. of nitric acid, the solution diluted 300 to 700 cc, according to the amount 
of copper taken, and then electrolyzed with a current of 1.5 to 2 amperes for 
thirty-six hours, with a spiral anode and a cathode with about 160 cm. depositing 
surface. The solution should remain slightly acid throughout the electrolysis, 
otherwise cobalt, nickel, and zinc may be precipitated as hydroxides from a 
neutral solution. When the copper is nearly removed, the electrodes are dis- 
connectedy and removed. 

The solution is concentrated by boiling, a few crystals of oxalic acid added, 
and the anode (which may be coated with Pb02) immersed in the hot solution 
for a few minutes, then rinsed off into the solution. 

Separation of Lead. The solution is evaporated to small volume, about 
40 cc. of dilute sulphuric acid (1:1) arc added and the mixture evaporated 
to S0» fmnes. The cooled concentrate is diluted with 100 cc. of water and 
again evaporated to fumes. About 300 cc. of water added and when the soluble 
salts have dissolved, the solution is filtered and the residue, PbS04, washed. 
The filtrate contains Zn, Ni, Co, etc. 

Determination of Lead. The residue, PbSO^, is dissolved by successive 
treatments with ammonium acetate and hot water, the lead precipitated from 
the solution, made slightly acid with acetic acid, by adding a slight excess of 
potassium chromate and the element determined as lead chromate according 
to the standard procedure for lead, b'ee page 236 in the chapter on Lead. 

Removal of the Hydrogen Sulphide Group. The filtrate from the lead 
sulphate is saturated with H2S and filtered. The filtrate contains zinc, cobalt, 
and nickel. To recover any occluded zinc, the precipitate is dissolved in nitric 
acid, taken to fumes with sulphuric acid, diluted to about 200 cc, and again 
treated with H2S. The filtrate from this precipitate is combined with the first 
portion. The precipitate is rejected. 

Removal of Iron. This, if present, will be found in the filtrate. The 
HjS is expelled by boiling and the solution concentrated to 400 cc. after adding 
5 cc. of H2O1 to oxidize the iron. Five grams of ammonium sulphate are added, 
the solution made strongly ammoniacal, and filtered. Iron is precipitated as 
Fe(OH)s and is thus removed. If much iron is present, a double precipitation 
is advisable to recover any occluded zinc, nickel, or cobalt, and the filtrates 
combined. 

Determination of Zinc. The filtrate from iron is concentrated to 400 
cc, then made neutral to litmus by cautious addition of dilute sulphuric acid, 
drop by drop, and then faintly acid with 3 drops in excess. Zinc is now pre- 
cipitated as the sulphide by saturating the solution with II2S and allowing 
to stand over night. The sulphide is filtered off. The filtrate contains cobalt 
and nickel. 

Zinc sulphide is dissolved in hot dilute HCl (1:2) and a few cr>'stals of 
KClOs. The solution is evaporated to dryness, the residue taken up water con- 
taining a few drops of HCl and the extract filtered. (To remove any SiOa dis- 



170 COPPER 

solved from the beakers.) Zinc carbonate is now precipitated (in a beaker of 
glass, which does not contain zinc) from the filtrate by addition of sodium carbon- 
ate, and ignited to the oxide ZnO. 

ZnO X 0.8034 =Zn. 

Detennination of Nickel and Cobalt. The filtrate from the zinc sulphide 
is examined for nickel and cobalt. About 0.5 cc. of sulphuric acid is added. 
US is expelled by boiling, and 2 cc. of HiOi added. The solution is concentrated 
to about 400 cc. (this should be free from nitric acid), treated with about 
25 cc. of ammonium hydroxide, and electrolyzed over night with a current of 0.5 
amperes. Nickel and cobalt, if present, are deposited on the cathode as metals 
and so determined. For greater details, consult the chapter on Nickel under 
the method by electrolysis. 

Determination of Arsenic, Antimony, Selenium, and Tellurium 

Separation of Copper. Ten to 50 grams of blister copper and 100 to 
500 grams of refined copper are required for the determination. (For 500-grams 
sample, 5 lots of 100 grams are taken.) The drillings are dissolved in nitric acid 
(40 cc. per 10 grams) and the solution boiled until a light-green precipitate 
appears on the surface. The liquor is diluted to 500 cc, and 5 cc. of ferric 
nitrate containing 3% of iron are added. A basic acetate precipitate is now made, 
weak sodium hydroxide being added to neutralize the free acid, but not in 
sufiicient amount to produce a permanent precipitate. If the end-point is 
ovemm, nitric acid is added drop by drop until the solution clears. The 
solution is diluted to about 800 cc, 20 cc. of a saturated solution of sodium 
acetate added, the liquor brought to boiling and filtered hot through a large 
creased filter paper, the first portion of the filtrate being poured back on the 
filter. The residue is washed twice with hot water to remove the copper. Five 
cc. additional iron are added to the filtrate and a second basic acetate precipitation 
made, a separate filter being used. The precipitates are dissolved in the least 
amount of nitric acid necessary and the solutions combined. The liquor is 
concentrated to 150 cc, a pinch of potassium chlorate added, and the con- 
centration continued until the volume has been reduced to about 30 cc. An 
equal volume of strong hydrochloric acid is added and a second pinch of chlorate 
and the evaporation repeated to eliminate all traces of nitric acid. 

The evaporation is best conducted in a casserole, resting in the circular 
opening of an asbestos board, in order that the sides of the vessel may be pro- 
tected from the flame. 

Separation and Determination of Arsenic. The solution is transferred 
to a distillation flask, arsenic reduced with ferrous chloride, and distilled according 
to the standard procedure for this element, p. 33. * In this distillate arsenic is 
determined volumetrically.* (See chapter on subject.) Antimony, selenium 
and tellurium remain in the flask. 

Separation and Determination of Antimony. Twenty-five cc of a saturated 
solution of zinc chloride are added to the liciuor remaining in the distilling 

*The concentration should not bo rarriod below 30 cc. 

' Arsenic may be precipitated by HjS, the sulphide dissolved in NH<OH, the filtrate 
taken to dryness, HNOi added and the evaporation repeated. Arsenic now is deter- 
mined by precipitation with AgNOj and titration of the silver with KCNS in presence 
of a ferric salt. AgX0.2316=A8. 



COPPER 171 

flask after the elimination of arsenic. The antimony is now distilled, strong 
hydrochloric acid being introduced in the distilling flask drop by drop by means 
of a separatory funnel, to replace the solution distilled, the volume in the flask 
being kept as low as possible, avoiding crystallization. When the antimony 
has been completely eliminated, the contents of the distilling flask is poured 
out while still hot, and, together with the rinsings of the flask, placed aside 
for the subsequent determination of selenium and tellurium. 

The distillate fa neutralized with ammonia, then made slightly acid with 
HCl and antimony precipitated with HjS. Most of the selenium and tellurium 
remain in the flask. Some ( f the selenium, however, dfatills with the antimony, 
hence thfa must be recovered from the antimony sulphide precipitate and at 
the same time thfa must be purified. 

The precipitate fa dissolved in dilute HCl (1 : 2), containing a little bromine 
to oxidize the sulphur. The solution is filtered free from sulphur and the filter 
washed with a little dilute HCl. The filtrate should contain one-third its volume 
of strong HCl. Selenium fa now precipitated by passing in SOj gas to satura- 
tion and bringing the solution to boiling. The precipitate is allowed to settle 
several hours and then filtered through a tared Gooch crucible. (To this fa 
added the selenium and tellurium later obtained from the residue of the flask.) 
The filtrate contains antimony. 

After boiling out the SOa, the filtrate fa first neutralized with ammonia, then 
made slightly acid with hydrochloric acid and antimony precipitated as the 
sulphide by saturating the solution with HjS, allowing the precipitate to settle, 
resaturating with HjS and again allowing to settle. The filtered, washed pre- 
cipitate fa dissolve<l with sodium sulphide, and 10 cc. of 25% potassium cyanide 
(poison) added to the filtrate, together with 2 cc. of 25% sodium hydroxide. 

The solution fa now electrolyzed hot (90*' C.) for an hour with a current of 
0.5 ampere and antimony deposited as the metal on the cathode. Thfa fa 
quickly removed and washed by dipping it successively into a beaker of cold 
water, three of hot water and one of 95% alcohol. The foil fa dried at lOO** C, 
and then weighed, on cooling, as usual. Antimony fa now removed by immersing 
the cathode in boiling nitric acid containing tartaric acid, and washing as 
before. The loss of weight of the foil fa taken as antimony. 

Note. It is advisable to test the electrolyte for antimony by acidifying the solu- 
tion with oxalic acid (Hood). A reddish coloration indicates the incomplt;te removal 
ot the element. 

Determination of Selenium and Tellurium. The solution from the dis- 
tillation flask fa nearly neutralized with ammonia and saturated with H2S. 
The precipitate fa filtered off and dissolved in equal parts of nitric acid (sp.gr. 
1.42) potassium bromide bromine solution (20 cc. Br added to a saturated 
solution of KBr and diluted to 200 cc). The liquor fa diluted to 400 cc, 5 cc 
of ferric nitrate (3% Fe'") solution added, and sufficient ammonia to make the 
solution decidedly alkaline. The precipitate contains, besides the iron, all of 
the selenium and tellurium, whereas any copper that may have been present 
fa removed. The precipitate, washed, fa dissolved in hydrochloric acid, the 
free acid nearly neutralized and H2S passed in to saturation. The precipitate 
fa filtered ofF, washed, and dissolved in the nitric acid potassium bromide and 
bromine mixture stated above. The solution fa filtered and then sufficient hydro- 
chloric acid added to make the solution contain about one- third its volume of 
strong HCl. Selenium and tellurium are precipitated from thfa solution by pass- 



172 COPPER 

ing in SO2 to saturation, and boiling for a minute or so. The precipitate is now 
filtered into the crucible containing the selenium obtained in the purification of 
the antimony precipitate. After washing with hot water and once with 95% 
alcohol, the residue is dried at 100° C, for an hour arid weighed as selenium and 
tellurium. Solution should stand three hours at least, or overnight, before filtering. 

Note. The precipitate of selenium and tellurium may contain gold, which should 
be determined by assay. 

Determination of Oxygen 

This determination is required only in refined copper. The method depends 
upon the reduction with hydrogen of cuprous oxide heated to redness; the water 
formed by the reaction being the measure of the oxygen. 

Apparatus. The combustion-furnace is the same as that used for the 
determination of carbon. As it is necessary that the hydrogen be absolutely 
free from oxygen and moisture, the gas is passed through a preheater con- 
sisting of a platinum or silica tube of small bore heated to redness by a flame 
or an electrical device. The gas is then passed through a tube containing 
calcium chloride and finally through a P2O5 bulb containing the anhydride. 
In this purified form it enters the combustion-tube. The product of com- 
bustion, water, is absorbed in a tared bulb by PjOs, to which is attached a tube 
of calcium chloride. 

Procedure. The sample, which has been drilled with considerable care 
to avoid overheating, is dried under partial vacuum in a desiccator after wanning 
to below 70** C. for a few minutes. 

One hundred grams are taken for analysis and placed in the combustion 
tube, the drillings being held in a large boat. Purified hydrogen is rapidly 
passed through the tube for half an hour to sweep out the air, the tube being 
cold. The tared PjOs bulb and the calcium chloride tube are now attached. 
The heat is turned on to bring the sample to cherrj*^ red heat, 900° C, and the 
current of hydrogen passed slowly over the sample for several hoiu^. 

The increase of weight of the PiO» bulb ==HiO. 

H,OX 0.8881 =0. 0X4.9687 =CuO. 



Determination of Sulphur 

This determination is rarely required in refined copper. 

Twenty grams of blister, unrefined or cement copper, placed in a casserole, 
are treated cold with 50 cc. bromine-potassiimx bromide mixture (see under 
Determination of Selenium and Tellurium). After standing at least ten 
minutes, 100 cc. of strong nitric acid are added. After another ten minutes 
the casserole is placed on the steam bath and the solution evaporated to small 
volume. This is taken up with 25 cc. of strong hydrochloric acid and evaporated 
to a pasty mass. The treatment is repeated to ensure tlie decomposition of 
nitrates and to expel nitric acid. It is now taken up with 5 cc. of hydrochloric 
acid, diluted with water and sulphuric acid precipitated as BaSO*, according 
to the standard procedure for sulphur. Sec p. 395. 

BaSO4X0.1374=S. 



COPPER 173 

Determination of Phosphorus 

This determination is seldom required, and then only in low-grade copper 
and copper scrap containing phosphor bronze. The sample, dissolved in nitric 
acid, is treated with ferric nitrate and the basic acetate precipitation made as 
has been described for the determination of arsenic, etc. The precipitate is 
dissolved in HCl, this solution then made strongly ammoniacal, and saturated 
with HiS, and filtered. The filtrate containing the arsenic and phosphoric 
acid is acidified, arsenic sulphide and sulphur filtered off, and phosphoric acid 
determined in the filtrate by precipitation with magnesia mixture as usual. 
See chapter on Phosphorus. 

Mg,P,OTX.2787=P. 

DETERMINATION OF COPPER IN REFINED COPPER 

' In determining the quality of copper for electrical purposes each hundredth 
of a percent above 99.90 has its significance. The methods employed are the elec- 
trolytic and the hydrogen reduction methods. Silver present is rated as copper. 

Electrolytic Method.^ The sample, consisting of unground drillings, should 
be untarnished, free of grease or oil, and cleaned of particles of iron by use of a 
good magnet. 

Procedure. A catch weight of about 5 grams is taken, each piece being ex- 
amined for dust, particles from the drill and surface oxidation before it is 
I^aced on the balance pan. Solution is effected in a special 400 cc. 
beaker which has hipped sides to support a series of watch 
glasses, the lower hip at the 125 cc. mark, the upper at 350 cc. 
(Fig. 32.) 

The drillings are treated with 50 cc. of a stock solution (10.5 
parts nitric acid and 4.5 parts of sulphuric). The watch-glass 
traps are put in place to retain the copper which is alwajrs 
entrained in the nitrogen peroxide fumes. Except that the cur- 
rent is maintained at .75 ampere throughout the period of elec- Fiq. 32. 
trolysis, the conditions are the same as have been described for 
the determination of copper by the " Small Portion Method." (Page 159.) 

Hydrogen Reduction Method. This method is applicable to the determina- 
tion of copper in grades of refined copiwr which are characterized by a metallic 
impurity content which is constant and less than 0.01 per cent. The apparatus 
consists of a combustion furnace, j^refcrably electrolytically heated, the tem- 
perature of which can be kept constant at about 050° C; a silica tube of |-in. 
bore, one end of which is connected witli a large Peligot tube containing con- 
centrated sulphuric acid, the other end is connected by a rubber plug and flexible 
tube with a source of purified hydrogen; porcelain combustion boats 95 mm. 
long, 18 mm. wide and 10 mm. deep. 

Procedure. A catch weight of about 25.1 grams of drillings is placed in the 
combustion boat, and the boat inserted in the silica tube. After passing hydrogen 
for half an hour through the cold tube, the temperature is raised to 950° C. and 
so maintained for two hours. If the furnace is of a type, which will permit the 
removal of the tube without disconnecting the train,' the tube is taken from the 

^ Ferguson, Jour. Ind. and Eng. Chem., May, 1910. 
* Electric Heating Apparatus Co., New York. 




174 COPPER 

funmce without interruption of the stream of hydrogen and cooled by a jet of cold 
air. Wheo cold, the mass of copper, the particles of which are cemented, is taken 
from the boat and weighed. 

Note. If the sample is allowed to become molten, the boat and tube will be coated 
with a film of copper. 

A convenient and efficient type of combustioD furnace, hinged design, is shown in 
Fig. 32b. This furnace may be purchased from the Electric Heating Apparatus Com- 
pany, New York City. 




Fio. 32a. — Combustion Furnace, Uineed Design, Type 70 — Shown with one "Spare" 

Uoit. Height to center, 9^'. 

Bj eouFUu' oi UiB EMinc Heailni APinntiM Com^tiif, New York CKt. 

CHLORINE IN CEMENT COPPER AND COPPER ORES 

If the material contains very little silver the following method is applicable 
in laboratories equipped with apparatus for furnace assaying. 

Ten grams of the finely ground sample placed in an 800 cc. beaker are treated 
with 600 cc. water, 100 cc. nitric acid (free from iodic acid) and the mixture 
brought to boiling by gentle heating. After filtration and thorough washing, the 
insoluble residue is treated repeatedly with additional water and acid, of the above 
proportion, until a test of the filtrate with silver nitrate indicates complete extrac- 
tion of the soluble chloride. The combined filtnites are treated with a slight excess 
of silver nitrate and chloride of silver precipitated and determined in the usual way. 
Page 124. 

On a separate 10 gram sample an assay of silver is made and the equivalent 
weight of chloride calculated. This equivalent is added to the weight of silver 
chloride obtained in the extract. The percent of chlorine is calculated from this 
result by the formula. 

Weight of AgCIX.2474xl00 
10 



-= gram chlorine. 



COPPER 175 



DETERMINATION OF COPPER IN BLUE VITROL 

This is best determined on a 2 gram sample of the finely powdered dry salt or 
a catch weight of approximately 2 grams if the salt is moist. Copper is deposited 
electrolytically, the electrolyte being diluted to 130 cc. and containing 4 cc. of 
nitric acid and 5 cc. saturated solution of ammonium nitrate. A current of .18 
amperes and an electrode of 11 J sq. in. depositing surface are used. If the salt 
contains insoluble matter consisting wholly of basic salts, complete solution is 
brought about by gently boiling after adding 4 cc. nitric acid and 25 cc. of hot 
water to the salt. If the insoluble matter shows a tendency to remain in suspen- 
sion, the presence of arsenic or antimony is indicated. In this case the impurities 
are precipitated along with ferric hydroxide as has been previously described 
under the notes on the electrolytic determination of copper in blister copper, 
page 162. 

DETERMINATION OF COPPER AND LEAD IN BRASS i 

One gram of the alloy is dissolved in 8 cc. nitric acid and the nitrous fumes 
are boiled ofF; if tin is present, 40 cc. of boiling water are added, the metastannic 
acid allowed to settle on the hot plate for fifteen minutes and filtered ofF. 
(Method for tin is accurate only for wrought brass; high iron or antimony 
interfere). 

The filtrate from the tin is electrolyzed for copper and lead. If the lead is 
less than 0.75 per cent, an ordinary sandblasted, spiral anode is used; if the 
amoimt of lead is 0.75 to 5 per cent a sandblasted gauze cylinder is necessary. 
For amounts of lead over 5 per cent either a smaller sample is taken or the greater 
part of the lead is precipitated as lead sulphate and the small amount of lead 
passing into the filtrate is recovered by electrolysis, using } ampere current per 
solution, after adding 3 cc. of nitric acid. For lead under 0.5 per cent; 5 cc. of 
1 : 1 sulphuric acid are stirred in, after the current has been passing for at least 
ten minutes. If the lead is high the sulphuric acid is added after the electrolysis 
has continued for at least an hour. Under these conditions no lead sulphate 
deposits from the solution and as long as the current passes, the sulphuric acid 
present does not attack the PbOa deposited. After the sulphuric acid is added 
the current is raised to i ampere per solution and the electrolysis continued over- 
night. 

The lead peroxide is dried at 250°C. for half an hour. The factor 86.43 
gives the equivalent per cent lead. (Factor determined from the average of a 
large number of tests made on pure lead. The factor is best obtained under the 
conditions of the laboratory where the determinations are made, as it varies slightly 
with change of conditions.) 

The copper on the cathode is washed, dried and weighed according to the usual 
standard procedure. 

^ Method of The National Brass and Copper Tube Company, communication by 
R. T. Roberts. 



FLUORINE 

Wilfred W. Scott 

F, at.wt. 19; D (air) 1.31^^\ sp.gr. (-187'') 1.14; m.p. -223i.b.p. ~187*C; 

acids, HF, HsSiF.. 

DETECTION 

Fluorine is the most active element known, and is by far the most active 
of the halogens, displacing chlorine, bromine, and iodine from their combinations. 
Etching Test. The procedure depends upon the corrosive action of hydro- 
fluoric acid on glass, the acid being Uberated from fluorides by means of hot 

concentrated sulphuric acid. This test is 
^ ^ appUcable to fluorides that are decomposed 

by sulphuric acid. The reactions taking place 
may be represented as follows: 

I. CaF,+H2S04=CaS04+2HF. 
II. Si02+4HF=2H,0+SiF4. 

The test may be carried out in the appa* 
ratus shown in the illustration. Fig. 33. A 
clear, polished glass plate 2 ins. square, free 
from scratches, is warmed and molten wax 
allowed to flow over one side of the plate, the 
excess of wax being drained off. A small mark 
is made through the wax, exposing the surface 
of the plate, care being exercised not to scratch 
the glass. If the test is to be quantitative. 
Board the marks should be of uniform length and 
E width. The powdered material is placed in a 
large platinum crucible (B) (a lead crucible 
will do) ; sufficient concentrated sulphuric acid 
is added to cover the sample. The plate (D) 
with the wax side down is placed over the 
Fig. 33.— Etching Test for Fluorine, crucible and pressed firmly down. To prevent 

the wax from melting, a condenser (C), with 
flowing water, cools the plate. An Erlenmeyer flask (C) is an effective and simple 
form of condenser, though a metalHc cylinder is a better conductor of heat. A 
little water placed on tlie plate makes better contact with the condenser. As a 
further protection a wide collar of asbestos board (E) may be placed as shown 
in the figure. In quantitative work, where a careful regulation of heat is nec- 
essary, the crucible is placed in a casserole with concentrated sulphuric acid or 
in a sand bath, containing a thermometer to register the temperature. The run 
is best conducted at a temperature of 200° C. (not over 210° — H2SO4 fumes). 

176 




A sbestos 



Cone H^SO^ 



kvjX-.T.ui'.r 



FLUORINE 



177 



After an hour the wax is removed with hot water and the plate wiped clean, 
and examined by reflected light for etching. A test is positive when the mark 
can be seen from both sides of the glass. Breathing over the etched surface 
intensifies the mark. 

Treatment of Fluo-Silicates not Attacked by Sulphuric Acid. The 
powdered material is mixed with about eight times its weight of sodium car- 
bonate and fused in a platinum crucible. The cooled melt is extracted with 
water. Calcium fluoride is thrown out from the filtrate, according to directions 
under Preparation and Solution of the Sample. The fluoride may now be 
tested as directed in the etching test or as follows by the hanging drop test. 

The Hanging Drop Test. The test depends upon the reaction SSiFi+SHjO 
=2H2SiF«+H2SiO,. 

If the material contains carbonates, it is calcined to expel carbon dioxide. 
Half a gram of the powdered dry material is mixed with 0.1 gram dried pre- 
cipitated siUca and placed in a test-tube, Fig. 34, about 
5 cm. long by 1 cm. in diameter. A one-hole rubber 
stopper fits in the txihe, A short glass tube, closed at the 
upper end, passes through the stopper extending about 
3 mm. below. Two or three drops of water are placed in 
this small tube by means of a i)ipette, nearly filling it. 
Two cc. of concentrated suli)huric acid are added to the 
sample in the test-tube and this immediately closed by 
inserting the stopper carrying the hanging drop tube, exer- 
cising care to avoid dislodging the drop of water. The 
test-tube is placed in a beaker of boiling water and kept 
there for thirty minutes. If an appreciable quantity of 
fluorine is present a heavy gelatinous ring of silicic acid 
will be found at the end of the hanging drop tube in the 
stopper. 

It is important to have material, test-tube, and rubber 
stopper dry, so that the deposition may occur as stated.^ 





Fig. 34. 

Hanging Drop Test 

for Fluorine. 



Note. Dr. Olsen^ makes the test by heating the sample in 
a small Erlenmeyer flask, with concentrated sulphuric acid. 
A watch-crystal with a drop of water suspended on its curved 
siu^ace is placed over the mouth of the flask. A spot etch is obtained in presence of 
fluorine. 

Black Filter Paper Test. According to Browning,' small amounts of fluorine 

may be detected by the converse method for detection of 
silicates and fluosilicates ( See silicon). The fluoride is 
placed with a suitable amount of silica, in a small lead cup, 
1 cm. in diameter and depth (Fig. 35) ; a few drops of 
concentrated sulphuric are added; the cup is covered by a 
flat piece of lead with a small hole in the center; upon the 
cover is placed a piece of moistened black filter paper and 
upon this a small pad of moistened filter paper. The cup 
is heated on the steam bath for ten or fifteen minutes. A white deposit will 

iC. D. Howard; Jour. Am. Chem. Soc, 1906, 28, 1238-1239. C. N., 1906, SO, 
420. 

* Communicated to the author by J. C. Olsen. 

» P. E. Browning, Am. Jour. Sci. (4), 32, 249. "Methods in Chemical Analysis," 
by F. A. Gooch. 



Fig. 36. 



178 FLUORINE 

be found on the under side of the black filter paper, over the opening in the 
cover, if fluorine is present in an appreciable amount. ro.OOl gram CaFi or 
above, and 0.005 gram NajAlFe wDl give the test.) 

ESTIMATION 

The determination of fluorine in the evaluation of minerals used for the 
production of hydrofluoric acid is of technical importance. The demand for 
ehmination of the use of fluorides for preservatives of food makes its estimation 
in small amounts of importance. 

Fluorine occurs only combined. It is found abundantly combined with lime 
in the mineral fluorspar, CaF2. It occurs as cryoUte, NasAlF*; apatite, 
3Ca8(P04)»CaFj. It is found in mineral springs, ashes of plants, in bones, and in 
the teeth (CaFj). It occurs sparingly, with aluminum and silicon, in topaz, 
and with cerium and yttrium in fluocerite, yttrocerite, also in wavelUte, wag- 
nerite. etc. 

Preparation and Solution of the Sample 

Fluorides of the alkalies, and of silver and mercury, are readily soluble; 
copper, lead, zinc, and iron fluorides are sparingly soluble; the alkaline earth 
fluorides dissolve in 100 cc. H2O as follows: BaFa =0.163 gram, SrF, =0.012 
gram, CaFj =0.0016 gram. 

Fluosilicates of potassium, sodium, and barium are shghtly soluble in water 
and practically insoluble if suflScient alcohol is added. 

Organic Substances.^ These are best decomposed by the lime method, the 
detaib of which are given in the chapter on chlorine under the section for the 
preparation and solution of the sample, p. 122. For fluorides in organic 
matter it is advisable to decompose the substance in a seamless nickel tube, 
40 mm. long by 4-5 mm. bore. The end of the tube is sealed with silver solder. 
The lime used should be soluble in acetic acid. The tube is heated to yellow 
heat for two hours. The lime is then extracted with acetic acid and fluorine 
determined as calcium fluoride. 

Silicious Ores and Slags. 0.5 to 1.0 gram of material is fused in a cru- 
cible with ten times its weight of sodium and potassium carbonates (1 : 1) and 
poured into an iron mould. If a porcelain crucible has been used, this is broken 
up and added to the cooled fusion. The mass is digested with about 200 cc. 
of hot water for an hour, the mass having been broken up into small lumps, 
(Kneeland recommends using an agate-ware casserole as diminishing the liability 
of subsequent bumping) * then boiled briskly for ten minutes longer and filtered, 
the solution being caught in a large beaker. The residue is washed with hot 
water, followed by a hot solution of ammonium carbonate and the insoluble 
material rejected. The silica is removed with ammonium carbonate, followed 
by the zinc oxide treatment of the second filtrate, as described under the section 
of Separations. In presence of appreciable amounts of fluorides, the gravi- 
metric precipitation of fluorine as calcium fluoride is recommended. 

» H. Meyer and A. Hub, Monatsch. ftir Chem., 1910, 31, 933-938. C. N., 1910, 
85,489. 

«E. Kneeland, Eng. and Min. Jour., 80, 1212. A. 11. Low, "Technical Methods 
of Ore Analysis." 



FLUORINE 179 

Caldum Fluoride.^ The product is best decomposed by fusion with sodium 
and potassium carbonates, after mixing the fluoride with 2.5 times as much silicic 
acid, followed by ten times its weight of carbonates. Most of the siUeic acid and 
all the fluorine will be changed to soluble alkali salts, while the calcium vnW be 
left as insoluble calcium chloride. The mixture should be heated gradually to 
prevent the contents of the crucible from running over the sides by a rapid evo- 
lution of carbon dioxide. The thin liquid fusion soon thickens to a pasty mass. 
The reaction is complete when there is no further evolution of carbon dioxide. 
The fused mass is now extracted with hot water as indicated above, and the 
soluble fluoride filtered from the calcimn carbonate residue. Silicic acid is 
removed from the filtrate by addition of ammonium carbonate. Traces of 
silicic acid are removed from the filtrate taken to near drjTiess, after neutraUzing 
the alkali with dilute hydrochloric acid (phenolphthalein indicator), by the zinc 
oxide emulsion method given under Separations. Fluorine is precipitated as 
calciimi fluoride, according to the procedure given later on page 180. 

Soluble Fluorides. The salts are dissolved in water. In presence of free 
acid a platinum dish should be used and the acid neutralized w^ith sodium car- 
bonate with addition of about one-fourth as much more in excess. The fluoride 
is then precipitated as calcium fluoride. 

Hydrofluoric Acid. The acid may be titrated with standard caustic. 
Determined gravimetrically, the acid is neutralized and fluorine precipitated as 
calcium fluoride or lead chlorofluoride (pages ISO, 181). 

Valuation of Fluorspar. Details of the procedure worked out by E. Bidtel, 
Chief Chemist, Fairview Fluorspar and Lead Company, are given at the close 
of the chapter. 

SEPARATIONS 

Removal of Silicic Acid from Fluorides. This separation is frequently 
required, especially in samples where the sodium and potassium carbonate 
fusion has been required for decomposition of fluosilicates, or calcium fluoride 
mixed with silicic acid. (See Preparation and Solution of the Sample.) 

To the alkaline solution about 5 to 10 grams of ammonium carbonate are 
added, the solution boiled for Ave minutes and allowed to stand in the cold 
for two or three hours. (Tread well and Hall recommend heating to 40° C, 
and allowing to stand over night.) The precipitate is filtered off and washed 
with ammonium carbonate solution. The fluoride passes into the filtrate, while 
practically all of the silicic acid remains on the filter. 

Small amounts of siUca in the filtrate are removed by evaporating the solu- 
tion to near dryness on the water bath, then neutralizing the carbonate with 
dilute hydrochloric acid (phenolphthalein indicator) added to the residue taken 
up with a little water. Upon i)oiliug the pink color is restored, the solution 
then cooled and acid again added to discharge the color; this is repeated until 
finally the addition of 1-2 cc. of 2 N. HCl is sufficient to discharge the color. 
Four to 5 cc. of ammoniaoal zinc oxide solution (moist ZnO dissolved in 
NH4OH — ^Low recommends 20 cc. of an emulsion of ZnO in NH40H) is added 
and the mixture boiled until ammonia has been completely expelled. The 
flrecipitate of zinc silicate and oxide is filtered pnd wa.she(l with water. The 
puoride is determined in the filtrate by precipitation with calcium cldoride 
as directed later. 

* Trcadwell and Hall, Analytical Chem., p. 472. 



180 FLUORINE 

Separation of Hydrofluoric and Phosphoric Acids. ' The method of Rose 
modified by Treadwell and Koch/ takes advantage of the fact that silver phos- 
phate is insoluble in water, whereas silver fluoride Is soluble. The alkaline 
solution of the salts of the acids (solution of the sodium carbonate fusions) is 
carefully neutralized with nitric acid and transferred to a 3(X)-cc. caHbrated 
flask. A slight excess of silver nitrate solution is added, and the mixture made 
to volume and thoroughly shaken. After settling, the solution is filtered 
through a dry filter, the first 10 to 15 cc. being rejected; 225 cc. of this filtrate 
is again transferred to a 300-cc. calibrated flask, the excess of silver precipitated 
by adding sodium chloride solution, and after diluting to the mark and shaking, 
the precipitate is again allowed to settle; 200 cc. of this solution is taken for 
analysis, after filtering as previously directed. This sample represents 50% of the 
original sample taken. Fluorine is now determined by one of the procedures 
outlined. 

Separation of Hydrofluoric and Hydrochloric Acids. The solution con- 
taining hydrofluoric and hydrochloric acids, in a platinum dish, is treated with 
nitric acid and silver nitrate. The chloride is precipitated . as the silver salt, 
whereas the fluorine remains in solution and may be filtered off through a glass 
funnel coated with paraffine or wax, or a hard rubber funnel. In presence of 
phosphoric acid, silver nitrate added to the solution will precipitate the phos- 
phate as well as the chloride, whereas the fluoride remains in solution. The 
phosphate may be dissolved out from the chloride by means of dilute nitric 
acid. 

Separation of Hydrofluoric and Boric Acids. An excess of calcium 
chloride is added to the boiling alkali salt solutions of the two acids. The 
precipitate is filtered off and washed with hot water. The residue, consisting of 
calcium fluoride, borate and carbonate, is gently ignited and then treated with 
dilute acetic acid, taken to dryness, and the residue taken up with acetic acid 
and water. Calcium acetate and borate are dissolved, whereas the fluoride 
remains insoluble and may be filtered off and determined. 



QRAVIMETRIC METHODS FOR THE DETERMINATION 

OF FLUORINE 

Precipitation as Calcium Fluoride 

The method utilizes the insolubility of calcium fluoride in dilute acetic acid 
in its separation from calcium carbonate, the presence of which facilitates 
filtration of the slimy fluoride. The reaction for precipitation is as follows: 

2NaF4-CaCl2 =CaF24-2NaCl. 

Procedure. Solution of the sample and the removal of silica having been 
accomplished according to procedures given under Preparation and Solution 
of the Sample, and Separations, the solution is neutralized, if acid, by the 
addition of sodium carbonate in slight excess; if basic, by addition of hydrochloric 
acid in excess, followed by sodium carbonate. To this solution, faintly basic, 
1 cc. of twice normal sodium carbonate reagent is added, followed by sufficient 

»Z. anal. Chem., 43, 469, 1904. "Analytical Chemistry," Vol. 2, by Tread- 
well and Hall. John Wiley and Sons. 



FLUORINE 181 

calcium chloride solution to precipitate completely the fluoride and the excess 
of carbonate, i.e., until no more precipitate forms, and then 2-3 cc. in excess. 
After the precipitate has settled, it is filtered and washed with hot water. (The 
filtrate should be tested for fluoride and carbonate with additional calcium 
chloride.) The precipitate of calcium fluoride and carbonate is dried and 
transferred to a platinum dish, the ash of the filter, burned separately, is added 
and the material ignited. After cooling, an excess of dilute acetic acid is added, 
and the mixture evaporated to dr>'ness on the water bath. The lime is con- 
verted to calcium acetate, while the fluoride remains unaffected. The residue 
is taken up with a little water, filtered and washed with small portions of hot 
water, by which procedure calcium acetate is removed, while calcium fluoride 
remains on the filter.* The residue is dried, separated from the filter and 
ignited. This, together with the ash of the filter, is weighed as calcium 
fluoride, CaF^. 

To confirm the result, the residue is treated with a slight excess of sulphuric 
acid and taken to fumes in a platinum dish. The adhering acid is removed as 
usual by heating with ammonium carbonate, and the ignited residue weighed as 
calcium sulphate. One gram of calcium fluoride should yield 1.7436 grams of 
calcium sulphate.^ 

CaO X 1.3924 =CaF2, or X0.677=F. 

Factors. CaF, X 0.4867 =^F, or X0.5126=HF, or X1.0757=NaF. CaSO* 
X 0.5735 =CaF,, or X 0.2937 =F, or X 0.2539 =HF. 

Precipitation of Fluorine as Lead Chlorofluoride 

The method, worked out by Starck,' takes advantage of the double halide 
formed by action of lead cliloride upon a soluble fluoride. The compound, 
PbFCl, is about fourteen times the weight of the fluorine it contains. Unfor- 
tunately, the compound is quite appreciably soluble in water,* so that a loss 
occurs if pure water is used for wasliing the precipitate. The method is Umited 
to the determination of soluble fluorides. 

The sample, made neutral, is treated with a large excess of a cold saturated 
solution (200 cc. PbCla per 0.1 gram NaF in 50 cc. solution) of lead chloride, 
the precipitate, settled over night, is filtered off in a weighed Gooch crucible, 
washed several times with a saturated solution of lead chlorofluoride, and 
finally two or three times with ice-cold water. The compound is dried two hours 
at 140-150** C, and weighed as PbFCl. 

*The results are slightly low, owing to the solubiUty of calcium fluoride: 
100 cc. H2O dissolves 0.0016 gram CaF^; 100 cc. 1.5 N. HCsHaO, dissolves 0.011 gram. 

•Low recommends disintegration of the fluoride with sulphuric acid, diluting 
the mixture with water, boiling with ammonium chloride, and then with ammonium 
hydroxide and hydrogen peroxide. Calcium oxalate is now precipitated from the 
filtrate and CaO detennined by titration with standard permanganate according to 
the usual procedure for determination of lime. 

» Z. anorg. Chem., 70 , 173 (1911); Chem. Abs., 5, 2049 (1911). 

*One hundred grams H2O at 18° C. dissolves 0.0325 gram PbFCl and 0.1081 
gram at 100** C. 



VOLUMETRIC METHODS FOR THE DETERMINATION 
OF FLUORINE 

Volumetric Determination of Fluorine — Formation of Silicon 
Tetrafluoride and Absorption of the Evolved Qas in Water. 
OfTerman's Method • 

Silicon tetrafluoride is formed by the action of Bulphuric acid upon a fluoride 
in presence of silica, tiia evolved gas is received in water and the resulting com~ 
pound titrated with standard potassium hydroxide. The following reactions 
take place: 

A. 3SiF*+2H,0=2H,SiF.+SiO,. 

B. H^iF,+6KOH=6KF+SiO,+4H,0. 

The method is suitable for determining fluorine in fluorspar in evaluation 
of this mineral. 

procedure. The powdered sample, containing the equivalent of 0.1-0.2 
gram calcium fluoride, is mixed with about three times its weight of pulverized 
quartz (previously ignited and kept in a desiccator), placed in the decom- 
position flask F, shown in Fig. 36, and about 1 gram of anhydrous copper sulphate 




added, followed by 25 cc. of concentrated sulphuric acid. The stopcock E 
is closed and the air bath heated gradually till in one-half hour the temperature 
haa risen to 220°. The cock B is now opened and air slowly forced through the 
apparatus (by means of water pump) at the rate of about three bubbles per 
second, the temperature being kept at 220°, and the flask containing the sample 
occasionally shaken. When the bubbles of silicon tetrafluoride have disappeared 
from F, the Same is removed, but the air current continued for half an hour 
longer. The solution in the receiving flask is now titrated with 0.1 N. KOH. 

Notes. The apparatus shown in the cut is the fonn recommended by Adolph, 
and the details of procedure are esaentially his. This uiclhod is preferred to that 
of PenReld,' in which an alcoholic solution of potasslunt chloride is used to absorb 
the tetrafluoride, and the liberated hydrochloric acid titrated with the standard 
alkali in presence of cochineal indicator. 

The results obtained by this method are generally low, but the procedure is use- 
ful for rapid valuation of tluors[)ar. 



'Z. angew. Chem., 3, 615 (1890). 
87, 11,2500(1915). 

» Am. Chem. Jour,, 1,27 (1879). 



, II. Adolph, Jour. Am. Chem. Soc., 



FLUORINE 183 

The bottles ii, B, C, and D are for the purpose of thoroughly dr\'inR the air, as 
moisture in the apparatus is to be avoided. G contains strong sulphuric acid, Ji 
m filled with glass beads to remove sulphuric acid spray. / and J are empty tubes. 
which should be thoroughly dry. The gas is completely absorbed in A'. 

Volumetric Determination of Fluorine — Ferric Cliloride Metliod ^ 

The procedure, worked out by (Irecf, depends upon the principle that 
a neutral aqueous solution of ferric chloride forms a white cr>'stalline precipi- 
tate with neutral solutions of alkali fluorides, the following reaction taking place: 

6NaF+FeCl, =Na,(FoF.)4-3XaCl. 

The double fluoride is only very slightly soluble in water and does not fonn the 
red compound Fe(CNS)» with sulphocyanates. The addition of sodium chloride 
makes the precipitation more complete. 

Procedure. Half a gram sample of the sodium or potassium salt is placeii 
in a 30Q-CC. Erlenmeyer flask, and dissolved in about 25 cc. of hot water, 
then cooled and 20 grams of sodium chloride and 5 cc. of })otas.sium sul])ho- 
c>'anate added (100 grams KCNS i)er 500 cc. H^O). The solution is titmted 
with a standard solution of ferric chloride (of such strength that 100 cc. is 
equivalent to about 1 gram of NaF) until a yellow color is produced. Ton cc. 
of alcohol and 10 cc. of ether are now added and the mixture shaken gently, 
then the flask closed and shaken vigorously. The titration is now continued 
until the ether layer is permanently colored red. 

Note. Commercial sodium fluoride frequently contains free hydrofluoric arid 
and silioo-fluoride. These are converted to the fluoride of scxiium bv titration with 
Bodiiun hydroxide in presence of phenolphthalcin to neutral reaction; the total fluoride 
may now be determined as described. 

The free acid may be determine by titrating the salt in an aqueous alcoholic solu- 
tion in presence of potassium chloride, which converts the silico-fluoride to the in- 
soluble potassium silico-fluoride. 

Colorimetric Determination of Fluorine — Method of Steiger- 

and Merwin ^ 

The method is based on the })leaching action of fluorine upon the yellow 
color produced by oxidizing a solution oif titanium with hydrogen i>eroxide. 
A known amount of titanium in solution is mixed with definite volume of the 
solution containing the fluorine and the tint compared with a standard soluticm 
containing an equivalent amount of titanium. Tiie extent of bleaching enables 
the computation of the fluorine j^resent. The method is applicable to deter- 
mination of fluorine in amounts ranping from 0.0(X)05 to 0.01 gram. Merwin 
has shown that large amounts of alkali sulf)hates have a bleaching action 
similar to fluorine. Addition of free acid, or rise of temi)erature, intensifies 
the color lost by bleaching. Aluniinum sulphate has no marked effect (m standard 
solutions, or on solutions bleached by alkali sulphates, but it restores the color 
to a considerable degree to solutions bleached by fluorine. Ferric sulphate luis 
a similar effect. Phosphoric acid bleaches a standard solution. Silica has little 

» Method by A. Grccf, Analyst . 1013, p. 521. C. N., 7, 3939 (1913). 
« G. Steiger, Jour. Am. Owm. Soc, 30, 219, 1908. 

»H. E. Merwin, Am. Jour. Sol. (4), 28, 119, 1909. Chem. Abs., 3, 2919 (1009). 
J. W. Millor, '* A Treatise on (Quantitative Inorganic Analysis." Chas. Grifiin & Co. 



184 FLUORINE 

effect. According to Merwin an accuracy of 0.002 gram may be expected, an 
error which is half that of the most reliable gravimetric method. 

Reagents. Standard Titanium Solution, An intimate mixture of 1 gram 
of TiOj and 3 grams of anunoniimx persulphate is heated until the vigorous 
action has ceased, and the anunonium sulphate is expelled. The residue is treated 
with 20 cc. of strong sulphuric acid, heated to fuming and, when cold, poured 
into about 800 cc. of cold water. When the suspended salt has dissolved, 
57.5 cc. of strong sulphuric acid are added, and the solution made up to 1000 
cc. (50 cc. or more of the solution should be analyzed for TiO»). One cc, will 
contain 0.001 gram Ti02. 

Standard Fluorine Solution. 2.21 grams of sodium fluoride, which has been 
purified by recrystallizing, washing, and igniting strongly, is dissolved in 1000 
cc. of water. One cc. will contain 0.001 gram fluorine. 

Sulphuric Acid. 95.5% solution, sp.gr., 1.84. 

Hydrogen Peroxide. Ordinary strength. 

Standard Colored Solution. The solution used in determining fluorine in 
materials fused with alkali carbonates contains 10 cc. of the titanium solution, 
4 cc. of hydrogen peroxide, and 4 cc. of concentrated sulphuric acid. 

Apparatus. Nessler Tubes 6 cm. long, 2.7 cc. in diameter are recommended 
by the authors. Colorimeters may be used in place of Nessler tubes. A very 
suitable type for this purpose is shown on page 245, Fig. 43. 

Procedure. Two grams of the powdered sample are fused with 8 grams 
of mixed sodium and potassium carbonates, the fusion taken up with hot water, 
and when leached, 3 to 4 grams of ammonium carbonate added. The mix is 
warmed for a few minutes and then heated on the water bath till the ammonium 
carbonate is decomposed and the bulk of liquid is small. Silica, ferric oxide, 
and alumina oxide are thrown down and are removed by filtration. The filtrate, 
which should not exceed 75 cc, is treated with 4 cc. of hydrogen peroxide, and 
then 10 cc. of standard titanium solution cautiously added (H2O2 prevents 
precipitation of TiO^ by the alkali carbonates), followed by 4 cc. of strong sulphuric 
acid to neutralize the alkali carbonates. The solution, neutral or slightly acid, 
acquires a light orange tint. A little sodium carbonate is added in just suf- 
ficient amount to discharge the color, and then a drop or so of acid to again 
restore it. The amount of excess acid now required depends upon the amount 
of fluorine present in the solution. For amounts of fluorine less than 0.0025 
gram (0.125% of sample), 3 cc. of acid are added. For amounts of 0.025 to 
0.012 gram fluorine, 12 cc. of acid are added. The solution is diluted to 100 cc. 

Comparison. The test solution is now compared with the standard solu- 
tion containing 10 cc. titanium reagent, and the same amount of acid and 
hydrogen peroxide as in the test sample, in a volume of 100 cc. If Nessler tubes 
are used, these are held over a white surface illuminated with diffused light. 
In the absence of a bleaching substance, such as fluorine, the two solutions will 
have the same tint, but in presence of fluorine the bleaching effect will cause 
the test solution to appear paler than the standard. The depths of the liquids 
are adjusted so that the tubes will have the same intensity of color when moved 
from right to left or reversed. Should the left eye pc^rceive a darker shade, the 
tube on the left will appear unifonnly darker whether it be the test sample or 
the standard. The comparative depths of the licjuids in the tubes are measured 
and the ratio obtained by dividing the depth of the fluorine solution by the 
depth of the standard and multiplying by 100. Reference may be made to the 



FLUORINE 



plotted curve shown in Fig. 37. The ratio 



Depth of F Sol. 
Depth of Standard 



















































































































































































































































































































































































































































^ 








































































S 






































































s 






















































g 00015 

1 


















s 






































































s 












































































































































s 




















































































































1 QOOlO 




















































































































































































































































s 













































































































































































































































































































































































































































































































































5 








T 











7 










6 











e 


b 








« 








9S 






lOO 



APPARENT PEBCEKTAGEOFTi(\ 
Fio. 37. 



, while the ordinate represents the amount of fluorine in the 2-grain 
sample. 

Example. Suppose the test solution =3.6 cm. and the standard =4.5 cm., 



0,0010 




^ 




ft 






1 






V^ 


l\ 


^ 


4s 


^-v 


■r 












\ 




w.^? 






-^ 












^[■%J<?^ 


^ 


fc 






^ 










fe r\r<^c 




\ 


N 




"'" 








1 


^ 


>>\ 


V. 






^ 


OD040 










i 


\ 




^ 







RATIO OFDEPTH50FCOLOR 
Fia. 38. 

the ratio then =80, from the curve it ia evident that the fluorine =0.00095 gram 
or 0.0475%, since a 2-gram sample was taken. 



186 FLUORINE 

According to Merwin, however, the bleaching effect of alkali sulphates, which 
are present, will make the ratio much higher than it would be if they were 
absent. (The sulphates alone give a ratio of 125.) This ratio should be deter- 
mined on two 8-gram portions of the alkali carbonate mixture used in the fusion 
and the correction made accordingly. K m = ratio of the blank thus obtained, 
and r the ratio of the final test, then the formula, according to Merwin, is 

— — ■=gram fluorine in the sample, 4 cc. excess sulphuric acid being used 

in the samples, or .-^^ = grams F, if 12 cc. of acid are used in testing larger 

DoCX) 

amounts of fluorine. The plotted curve. Fig. 38, is that given by Merwin, 

and shows the effect of acidity on the depth of color obtained. The abscissa 

represents the ratio of the solutions, and the ordinate the amount of fluorine 

in grams. Temperature of the tests was 22° C. 

VALUATION OF FLUORSPAR 

The following procedure, worked out by Dr. Bidtel,* meets the commercial 
requirements for the valuation of fluorspar. The determinations usually required 
are calciimi fluoride, silica, and calcium carbonate; in some particular cases 
lead, iron, zinc, and sulphur. 

Procedure. Calcium Carbonate. One gram of the finely powdered sample 
is placed in a small Erlenmeyer flask, 10 cc. of 10% acetic acid are added, a 
short-stemmed funnel inserted in the neck of the flask as a splash trap, and 
the mixture heated for an hour on a water bath, agitating from time to time. 
The calcium carbonate is decomposed and may be dissolved out as the soluble 
acetate, whereas the fluoride and silica are practically unaffected. The solution 
is filtered through a 7-cm. ashless filter, the residue washed with warm water 
four times, and the filter burned off in a weighed platinimi crucible at as low 
a temperature as possible. The loss of weight minus 0.0015 gram (the amount 
of calcium fluoride soluble in acetic acid under the conditions named) is reported 
as calcium carbonate. 

Silica. The residue in the platinum crucible is mixed with about 1 gram 
of yellow mercuric oxide, in form of emulsion in water (to oxidize any sulphide 
that may be present); any hard lumps that may have formed are broken up, 
the mixture evaporated to dryness and heated to dull redness, then cooled 
and weighed. About 2 cc. of hydrofluoric acid are added and the mixture 
evaporated to dryness. This is repeated twice to ensure complete expulsion 
of silica (as SiFi). A few drops of hydrofluoric acid are then added, together 
with some macerated filter paper, and a few drops of ammonium hydroxide to 
precipitate the iron. The solution is evaporated to dryness, heated to dull 
redness, cooled and weighed. The loss of weight is reported as silica. 

Calcium Fluoride. The residue is treated with 2 cc. of hydrofluoric acid 
and 10 drops of nitric acid (to decompose the oxides), the crucible covered and 
placed on a moderately warm water bath for thirty minutes, the lid then 
removed and the sample taken to dryness. The evaporation with hydrofluoric 
acid is repeated to ensure the transposition of the nitrates to fluorides, and if 

* Dr. E. Bidtel, Chemist, Fairview Fhiorspar and Lead Company, Jour. Ind. Eng. 
Chem., Vol. 4, No. 3, March, 1912. 



FLUORINE 187 

the residue is still colored, hydrofluoric acid again added and the mixture taken 
to dryness a third time; then a few drops of hydrofluoric acid are added and 
10 cc. of ammoniiun acetate solution (the acetate solution is made by neutral- 
izing 400 cc. of 80% acetic acid with strong ammonia, adding 20 grams of citric 
acid and making the mixture up to 1000 cc. with strong ammonium hydroxide). 
The mixture is digested for thirty minutes on a boiling water bath, then filtered 
and washed with hot water containing a small amoimt of ammoniimi acetate, 
and finally with pure hot water. (Several washings by decantation are advis- 
able.) The residue is ignited in the same crucible and weighed as calcium fluoride. 
An addition of 0.0022 gram should be made to compensate for loss of CaF2. 

Pure calcium fluoride is white. To test the purity of the residue, 2 cc. of 
sulphuric acid are added and the material taken to fumes to decompose the 
fluoride; 1 cc. of additional sulphuric acid is added and the excess of acid 
expelled by heating. The residue is weighed as calcium sulphate. This is now 
fused with sodium carbonate, and the fusion treated with hydrochloric acid 
in excess. If barium is present the solution will be cloudy ( «BaS04.) 

ANALYSIS OF SODIUM FLUORIDE 

Preparation of the Sample and Insoluble Residue. Ten grams of the 
sample are dissolved in 250 cc. of water in a beaker, and boiled for five minutes, 
then filtered into a liter flask through an ashless filter; the residue is washed with 
several portions of water and ignited. This is weighed as insoluble residue. 
The filtrate and washings are made to 1000 cc. with distilled water. 

Sodium Fluoride. Fifty cc. of the solution equivalent to 0.5 gram of 
sample are diluted to 200 cc. in a beaker, 0.5 gram sodiimi carbonate is added 
and the mixture boiled. An excess of calciiun chloride solution is now added 
slowly and boiled for about five minutes. A small amoimt of paper pulp is 
added to prevent the precipitate from running through the filter, the precipitate 
allowed to settle and then filtered, using a 9-cm. S. & S. 590, or B. & A. 
grade A, filter paper. The fluoride is washed twice by decantation, and four 
or five times on the filter with small portions of hot water. The final washings 
should be practically free of chlorine. 

The residue is ignited in a platiniun dish, then treated with 25 cc. of acetic 
acid, and taken to dryness. This treatment is repeated and the residue taken 
up with a little hot water and filtered. The calcium fluoride is washed free of 
calcium acetate with small portions of water, remembering that CaFt is slightly 
soluble in water. The ignited residue is weighed as CaFi. 

CaFaX 1.0757 =NaF. 

Sodium Sulphate. To the filtrate from calcium fluoride is added 10 cc. 
hydrochloric acid and then a hot solution of bariimi chloride. The BaS04 is 
allowed to settle, filtered, washed, dried, ignited, and weighed as usual. 

BaS04X 0.6086 =Na,S04. 

Sodium Carbonate. Sodium carbonate is determined on a 5-gram sample 
by the usual method for carbon dioxide as described in the chapter on Carbon. 

Approximate results may be obtained by adding a small excess of normal 
sulphuric acid to 5 grams of ihe fluoride in a platinum dish, boiling off the carbon 



188 FLUORINE 

dioxide, and titrating the excess of acid with normal caustic, using phenolphthalein 
indicator. 

One cc. N. H2SO4 =0.053 gram Na,CO,. 

H,SO4Xl.0816=Na,CO,. 

Sodium Chloride. Fifty cc. of the sample is titrated with N/10 AgNOj 
solution. 

Silica. This is probably present as sodium fluoride and silicate. One gram 
of the sample is dissolved in the least amount of water and a small excess of 
hydrofluoric acid added to convert the silicate to silico-fluoride, then an equal 
volume of alcohol. After allowing to stand for an hour, the precipitate is filtered, 
washed with 50% alcohol until free of acid and the filter and fluoride are placed 
in a beaker with 100 cc. of water, boiled and titrated with N/10 NaOH. 

One cc. N/lONaOH =0.0015 gram SiO, or 0.0047 gram Na,SiF«. 

Volatile Matter and Moisture. One-gram sample is heated to dull redness 
to constant weight. Loss of weight is due to moisture and volatile products. 



DETERMINATION OF TRACES OF FLUORINE 

An approximate estimation of traces of fluorine may be made by utilizing 
the method outlined for detection of this element. The apparatus ^ is the 
same, with the exception that the crucible rests in a paraffine * bath containing 
a thermometer to regulate the temperature. A casserole may be used to hold 
the parafiine. By varying the amounts of substance tested an etch is obtained 
that is comparable with one of a set of standard etches, obtained with known 
amounts of fluorine in form of calcium fluoride, added to the same class of material 
examined. 

The conditions in obtaining the standard etches and those of the tests should 
be the same. This applies to the temperature of the paraffine bath, duration 
of the run, size of mark exposing the surface of the test-plate, and the general 
mode of procedure. 

Note. The importance of regulating the temperature may be seen by the results 
obtained by Woodman and Talbot. With a temperature of 79-82° C., one part 
of fluorine may be detected in 25 to 100 thousand parta of material; by raising the 
temperature to 136° C, the delicacy of the procedure is increased to one part of 
fluorine in 1 to 5 million parts. The limit of delicacy is apparently reached at 213-218° 
C. (i.e., 1 part F per 25 million). 

* A metal condenser, such as is recommended for mercurj' determinations, may 
be used and the paraffine bath substituted by an electric heater automatically con- 
trolled. 

*Crisco is claimed to be better than paraffine, as this does not give off any un- 
pleasant fumes when heated. 



GLUGINUM (BERYLLIUM) 

W. W. Scarr. 
Gl, at.wt. 9.1; sp.gr. 1.85^; m,p. > 960** C; oxide, GIO. 

DETECTION 

In the usual course of analysis glucinum will be precipitated by ammonia along 
with iron and aluminum hydroxides. Silica having been removed by evaporation 
to dryness of the acid solution of the substance, extraction of the residue with 
dilute hydrochloric acid and subsequent filtration; the members of the hydrogen 
sulphide group are precipitated from sUghtly acid solution by hydrogen sul- 
phide. The filtrate is concentrated to about 30 cc, and about 2 grams of sodium 
peroxide are added to the cooled liquid, which is now heated to boiling and 
filtered. Fe(OH)j remains insoluble, if iron is present, while aluminum and 
glucinum dissolve. The filtrate is acidified with nitric acid, and ammonia then 
added in excess. If a precipitate forms, alumina or glucinum or both are indi- 
cated. Glucinum hydroxide and aluminum hydroxide cannot be distinguished 
by appearance; the solubiUty of the former in sodium bicarbonate solution makes 
it possible to separate the two. The precipitate is dissolved in acid and the 
solution made almost neutral with ammonia. Solid sodium bicarbonate is added 
in sufficient amount to make the solution contain 10% of the reagent and the 
mixture heated to boiling, then filtered. Alumina hydroxide remains on the filter 
paper and glucinimi passes into the filtrate, in which it may be detected by 
diluting to ten volumes with water and boiling, whereupon glucinum hydroxide 
precipitates. 

Glucinum hydroxide, Gl(OH)j, is precipitated from neutral or acid solu- 
tion by ammonia, insoluble in excess (distinction from Al(OH)a). It is pre- 
cipitated by sodium and potassium hydroxides, soluble in excess (separation 
from iron); if this solution is boiled Gl(OH)i is reprecipitated, A1(0H)3 remains 
in solution. Gl(OH)i is soluble in an excess of ammonium carbonate, Al (0H)3 
is insoluble. 

ESTIMATION 

Gluciniun occurs in the minerals beryl, euclase, davalite, chrysoberyl, helvite, 
leucophane, phencaite. 

The oxide, GIO, is soluble in strong sulphuric acid. It is decomposed by 
fusion with potassium fluoride. The freshly precipitated hydroxide, G1(0H)2, 
is easily soluble in dilute acids, in alkalies and alkali carbonates and bicar- 
bonates. 

The methods of preparation and solution of the sample are the same as 
those described for the estimation of aluminum. For details of these procedures 
the analyst is referred to the chapter on this element. 

189 



190 aLUCINUM (BERYLLIUM) 

SEPARATIONS 

Removal of Silica and Members of the Hydrogen Sulphide Group. See 

proceduie given under " Detection." 

Separation of Glucinum from Iron and Manganese. The acid solution 
is nearly neutralized with ammonia and then poured with constant stirring into 
an excess of a cold mixture of ammonium sulphide and carbonate. Iron and 
manganese are precipitated, whereas glucinum passes into the filtrate. (Zir- 
conium and yttrium will be found with glucinum, if they are present in the 
material examined.) 

Separation from Zirconium and Yttrium. The filtrate obtained from 
the separation of iron and manganese is boiled for an hour, the precipitate 
is filtered and washed, then dissolved in dilute hydrochloric acid. To this 
solution is added an excess of sodium hydroxide, zirconium and yttrium are pre- 
cipitated, whereas glucinum remains in solution. After filtering, glucinum 
may be precipitated by boiling the diluted filtrate. 

Separation from Aluminum, Chromium and Iron. The elements precip- 
itated as hydroxides are ignited to oxides and fused with sodium carbonate 
for an hour or more. Upon leaching with water, aluminum and chromium dis- 
solve, while iron and glucinum remain insoluble. The oxides of glucinum and 
iron may be separated by fusion with sodium acid sulphate, extracting with water 
and precipitating the iron with an excess of sodium hydroxide, glucinum re- 
maining in solution. 

Separation of Glucinum from Aluminum. The hydroxides of alumina 
and glucinum are precipitated with ammonia and the precipitate treated with 
an excess of ammonium carbonate. G1(0H)2 dissolves, whereas Al(OH)i re- 
mains insoluble. See Detection, also Gravimetric Method for Determination 
of Glucinum. 



GRAVIMETRIC DETERMINATION OF GLUCINUM 

The procedure recommended by Parsons and Barnes » depends upon the 
solubility of glucinum hydroxide in a 10% sodium bicarbonate solution, in the 
separation of this element from iron and aluminum hydroxide precipitate, with 
which it is commonly thrown out from solution. (Uranium, if present, also 
dissolves.) 

Procedure. Silica and the members of the hydrogen sulphide group having 
been removed by the usual methods (See Detection), hydrogen sulphide is 
expelled by boiling, nitric acid is added in sufficient amount to oxidize iron 
(the hydrochloric acid solution turns yellow) and ammonium hydroxide added 
in slight excess. The precipitated hydroxides are allowed to coagulate by heating 
to boiling and, after settling a few minutes, filtered and washed with a 2% 
solution of anunonium acetate containing free ammonia. 

Separation from Iron and Aluminum Hydroxide. The precipitate is 
dissolved in hydrochloric acid, the solution oxidized with nitric acid or hydro- 
gen peroxide (C.P.), if necessary, and the free acid then neutralized with ammonia. 
To the cold solution are added 10 grams of sodium bicarbonate for each 100 cc. 

'C. L. Parsons and S. K. Barnes, Jour. Am. Chem. See, 28, 1589, 1906. 



GLUCINUM (BERYLLIUM) 191 

of liquid. The mixture is heated to boiling and boiled for one minute/ then 
cooled and filtered. The residue is washed with hot 10% solution of sodium 
bicarbonate. Iron and aluminum hydroxides remain on the filter and gluci- 
num passes into the filtrate. 

To recover occluded glucinum from the hydroxides of iron and alumina, 
the precipitate is dissolved in a few drops of hydrochloric acid, and the pre- 
cipitation repeated. It is advisable to repeat this treatment a third time, 
adding the filtrates to the first portion containing the glucinum. 

Precipitatioii of Glucinum. The combined filtrates from the almnina and 
iron hydroxides are acidified with strong hydrochloric acid, the beakers covered 
to prevent loss by spurting and the carbon dioxide completely removed by 
boiling. (COi remaining in solution would form anmionium carbonate, on sub- 
sequent treatment with ammonia, which would dissolve glucinum.) A slight 
excess of ammonia is hdw added, the mixture again boiled and the precipitated 
glucinum hydroxide allowed to settle, then filtered and washed with a 2% solu- 
tion of ammonium acetate containing free ammonia, until the chlorides are removed. 
After ignition the residue is weighed as glucinum oxide, GIO. 

GIO X 0.3626 =G1. 

* Prolonged boiling would cause the loss of too much CO2; so that Al(OH)i would 
be apt to pass into solution. The evolution of COs may be mistaken for boiling. 



GOLD 

W. G. Derby 
Au, at.wt. 197.2; sp.gr. 19.33; nup. 1063; 6.p. 2530'' C; oxides, Au^O, AUjOi 

DETECTION 

Because of the limited application and tediousness of wet methods, the 
detection of a small quantity (2 parts i)er million or less) of gold in a mineral or 
base met 1 is most positively carried out by furnace methods of assaying. 
Wet methods of detection of traces of gold can be applied only to solutions free of 
colored salts and elements precipitated by the reagents employed. As a rule, 
in the treatment of an unknown substance, advantage is taken of the solubility 
of most metals and their compounds, and insolubility of gold by one of the mineral 
acids. 

Detection of Gold in Alloys. In metals or alloys which produce colorless 
solutions with dilute nitric acid, gold, in the absence of other insoluble matter, 
exhibits itself as a black or brownish residue which settles readily, and from which 
the liquid can be separated by careful decantation. If unassociated with metals 
of the platinum group, this residue will become yellowish brown on heating with 
strong nitric acid. 

In copper, nickel and such alloys, which leave a residue of sulphur, carbon or 
silicious matter on treatment with dilute nitric acid, the solution is filtered through 
double ashless filters and the filter and residue incinerated in a porcelain crucible. 
The residue, which may require pulverizing, is digested for a few minutes with 
aqua regia, and the dilute, filtered solution evaporated to dryness by heating 
below 200° F. Just as soon as dry, the mass is moistened with the least quantity 
of hydrochloric acid and the purple of Cassius test applied to its water solution 
in a small volume. This test is made by adding a solution of stannous chloride, 
containing stannic chloride. In strongly acid and concentrated gold solu- 
tions a precipitate of brown metallic gold is obtained. If the solution is but 
slightly acid and dilute, a reddish purple color is produced by colloidal gold and 
the stannic acid. The tint fades on standing. Addition of ammonia produces 
a red coloration. 

This test applied to 1 part of gold in 600,000 of solution will impart a per- 
ceptible shade; to double this quantity, a mauve color. When gold is present in 
somewhat greater proportion a flocculent precipitate will form. 

Test for Gold in Minerals. From minerals, in which the metal exists in unal- 
loyed, or uncombined state, gold may be extracted by iodine in potassium iodide 
solution, or by chlorine or bromine water. All minerals containing sulphides 
should be roasted. In natural or roasted state the sample should be ve y finely 
pulverized, and usually yields the gold best if first digested with nitric acid and 
washed free of soluble salts. The sample in a flask is covered with bromine 
water, the flask closed with a plug and shaken frequently during a period of three 

192 



GOLD 193 

or four hours. The purple of Cassius test is applied to the extract, removed by 
decantati )n after con entration. 

If it is evident that base metals are present in the bromine water extract in 
quantity sufficient to mask the purple of Cassius test, hydrogen peroxide is added 
to the concentrated liquid, slightly alkaline with sodium or potassium hydroxide 
or carbonate.^ After boiling the solution until hydrogen peroxide is removed, 
precipitated hydroxides or carbonates are dissolved by hydrochloric acid. Gold in 
exceedingly small quantity exhibits itself as a light-brown residue on a fine filter. 
This indication should be confirmed by a purple of Cassius test on the aqua regia 
solution of the residue; the test carried out in the same manner as on the residue 
from a solution of a metal. 

Benzidine Acetate Tests. Maletesta and Nola * make use of benzidine acetate 
(1 gram benzidine dissolved in 10 cc. acetic acid and 50 cc. water) as a reagent in 
the detection of gold and platinum in quite dilute solutions. Gold gives a blue 
coloration which gradually changes to violet. The coloration is green in the 
presence of free acetic acid, changing to blue with addition of benzidine in excess. 
Platinum gives a blue flocculent precipitate, the formation of which is pro- 
moted by heating. Free mineral acids have no influence on the gold and retard 
the platinum reaction only in the cold. Since ferric salts give a blue colora- 
tion, stable only in excess of benzidine, their absence must be assured before 
application of the test for the precious metals. The limit of sensitiveness 
of the test is 35 parts for gold and 125 parts for platinum per 10,000,000. 

Phenylhydrazine Acetate Test. E. Pozzi Escot' adds phenylhydrazine 
acetate to a very dilute gold solution which contains an excess of an organic acid 
(formic or citric). A violet coloration, permanent for several hours, is imparted. 
The depth of color is proportional to the quantity when the gold is present in less 
amount than one part in 500,000. 

ESTIMATION 
Solubility 

Gold in massive form is practically insoluble in pure nitric, sulphuric or hydro- 
chloric acids, but in the presence of oxidizing agents, is attacked appreciably by 
sulphuric, and actively by hydrochloric acid. Gold is found in minute quantity 
in the nitric acid* solution of its alloys and in such as contain selenium, the 
amount may be a large part of the total present. 

Gold is attacked energetically by aqua regia. Large amounts of gold are 
dissolved with requirement of least attention when the proportion of hydrochloric 
acid is several times that of the aqua regia formula, (3HC1 : IHNO3). 

Gold is dissolved by solutions of chlorine or bromine, by alkaline thiosulphates; 
in the presence of free oxygen by iodine in potassium iodide solution, by soluble 
cyanides, by fused potassium or sodium hydroxide; by fused potassium or sodium 
nitrate or sulphide. In a finely divided state, it is dissolved by a solution of potas- 
sium or sodium hydroxide. 

Gold alloys quickly with molten lead. When in the form of bright, untarnished 
particles it alloys readily with mercury. 

' Vanino and Seeman, Berichte, 32, 1968; Rossler, Zeit. Anal. Chem., 49, 733. 

« Bull. Chim. Farm, 62, 461; Chem. Abs., April 20, 1397, 1914. 

» Am. Chim. Anal. AppL, 1007, 12, 90; J.S.C.I., June 15, 1907, 645. 

* Dewey, J.A.C.S., March, 1910, 318; E. Keller, Bull. Am. Inst. Min. Eng., 67, 681. 



194 GOLD 

GRAVIMETRIC METHODS 

Gold is alwa3rs weighed in metallic state, and is determined most accurately 
in the form of^the mass obtained by dilute nitric acid treatment of the silver 
alloy resulting from the operation of cupellation in the method of assaying by 
furnace processes. On account of tediousness in making complete separation 
from associated metals, and of uncertainty in collection of the product in a 
form suitable for accurate weighing, direct precipitation methods are never used 
for the valuation of gold-bearing material, but may be applied to the estimation 
of gold in plating baths, the Wohlwill parting electrolyte and solutions of similar 
type. 

Precipitation of Gold. From such solutions of auric chloride, slightly acid 
with hydrochloric, freed of oxidizing agents by evaporation and displacement with 
hydrochloric acid, and containing but little of the salts of the alkalis or alkali 
earths, gold is separated from other than occluded platinum and palladium by 
precipitation with oxalic acid, ferrous sulphate, or hydrazine hydrochloride. The 
reactions are hastened by heat. When salts of the alkalis or earths are present, 
equally good separation and more complete precipitation can be obtained 
by addition of excess of sodium peroxide, boiling vigorously for a few minutes 
and then acidifying with hydrochloric acid. The precipitated metal is collected 
on an ashless filter paper, and after drying, weighed. 

Gold precipitated from a very weak solution is in such fine form that it is not 
wholly retained by the finest paper. 

Wet Qold Assay of Minerals 

A wet gold assay, suitable for prospector's use,^ is carried out by covering one 
assay ton (29.17 grams), of the finely pulverized natural or roasted ore, in a por- 
celain mortar, with 50 cc. of a solution of 2 parts of iodine and 4 parts potassium 
iodide in 100 cc. of water. Sulphide ores should be roasted and digested with 
nitric acid before treatment with the iodine solution. Simila;* treatment is 
advantageously applied to all ores. The ore is ground in contact with the iodine 
solution and additions of the halogen are made whenever the liquid becomes color- 
less. The solution is then allowed to stand at least an hour. To the filtrate 
and washings from the pulp, in a glass-stoppered bottle or flask, are added 5 
grams of gold free mercury. The liquid is shaken vigorously with the mercury 
until clear. The mercury is then transferred to a small porcelain casserole, 
washed with clean water and dissolved by warming carefully with 10 cc. nitric 
acid. The gold mass is washed free of nitrate of mercury by decantation, dried 
and annealed by heating in a casserole over a Bunsen flame, and the metal 
weighed. Each miUigram represents an ounce per ton. Results obtained by this 
method of assaying are usually more than 50 per cent of the actual gold content. 

£lectrol]rtic Method. The gold content of a cyanide plating bath containing 
no potassium ferrocyanide may be estimated by electrolysis.* 

Procedure. A measured quantity, 25 to 50 cc. in a tared platinum dish, is 
diluted to 1 cm. of the rim of the dish and using a carbon or platinum anode, elec- 

» De Luce, Min. Sci. Press, 100. 895; Hawson, Min. Sci. Press, 100, 936; Davis, 
Mines and Minerals, Oct., 1910, I*eb., 1911; Austen, Inst, of Min. and Met., May 
31, 1911. 

* Electro Deposition of Metals, Langbein. 



GOLD 195 

trolyzed for about three hours at a current density NDioo =0.067 amp. (.0.0043 per 
square inch). Completion of deposition is recognized by the lack of any deposit 
within fifteen minutes, on a platinum strip suspended on the rim of the dish. The 
dish plus gold deposit is washed, rinsed with dcohol, dried at 212° and when cold 
weighed. 

The following is a summary of the conditions of deposition of gold in compact 
form as described by Classen * 3 grams potassium cyanide were added to a gold chloride 
solution containing 0.0545 grams of gold m 120 cc. This solution heated to about 55° C. 
when electrolyzed at a current density of NDioo= 0.38 amp. (0.024 amp. per square inch) 
with a potential difference of 2.7-4.0 volts, deposited its gold content in one and a 
half hours. Time required for deposition is tripled if the electrolyte is at room tem- 
perature. 

Miller' deposited 0.1236 gp*am of gold in two and a quarter hours from 125 cc. 
of electrol3rte at 50° C. contamin^ 1 gram potassium cyanide by a current of NDioo= 
0.03 amp. (0.002 amp. per square mph) and 2.5 volts. 

Perkm and Preble ' use an electrolyte containing ammonium thiocyanate in place of 
potassium or sodium cyanide. 

Gold is removed from the platinum electrode by wanning with a solution of chromic 
anhydride in a saturated salt solution/ or with a solution of potassium cyanide con- 
taining some oxidizing agent as hydrogen peroxide, sodium peroxide or alksdi per- 
sulphate.* 

VOLUMETRIC METHODS 

These methods are applicable to the determination of the strength of chloride 
of gold solutions used in photography, electro gilding, and as electrolyte in the 
Wohlwill parting process. 

Preparation of the Sample. Nitric acid or nitrates in the solutions should be 
removed by repeated evaporations to syrup with addition of hydrochloric acid 
saturated with chlorine. Free chlorine or bromine should be removed by addition 
of ammonia to formation of permanent precipitate, then making the solution 
very slightly acid with hydrochloric acid and heating until the precipitate of 
fulminating gold dissolves. The gold solution should contain but little free 
hydrochloric acid, an excessive amount of which may be removed by ammonia. 

Permanganate Method 

Weak gold solutions should be concentrated whenever possible. The perman- 
ganate method,* which is not applicable when the sample contains organic matter, 
depends upon the titration, after complete precipitation of gold, of the unoxidized 
portion of a measured quantity of an added reagent of a known gold precipi- 
tating value. The reagent may be ammonium or potassium oxalate, ferrous 
sulphate or ferrous ammonium sulphate in solutions varying from 5 to 25 milligrams 
gold precipitating value and is titrated with a permanganate solution of approx- 
imately equal oxidizing strength. One part of gold requires for precipitation 1.08 
of ammonium oxalate, 1.40 of potassium oxalate, 4.22 of ferrous sulphate, 5.96 

^Classen, "Quantitative Chemical Analysis by Electricity," Classen-Boltwood. 
« J. A.C.S., Oct., 1904, 1255. 

• Elec. Chem. and Met. Ind., 3, 490. 

• Classen-Boltwood, " Quantitative Chemical Analysis by Electricity." 
» Rose, " Met. of Gold," 5th Ed., 469. 

• Bull. Chim. Farmac, 1894, XXX, III, 35; Oestr. Zeit. f. Berg, und Hut.. 182, 1880: 
Sutton, " Volumetric Analysis," 10th Ed.; £. A. Smith, " Sampling and Assaying of 
Precious Metals "; Min. Eng. World, 87» 853. 



196 GOLD 

parts ferrous ammonium sulphate, each in crystalline form. The most satis- 
factory precipitations are made with the iron salts. The standard solution of 
either should contain about 0.1 per cent of sulphuric acid. One part of gold, in 
solution as auric chloride, has an oxidizing value equivalent to 0.4808 part of 
potassium permanganate. 

The precipitating value of 0.2548 gram of dry Sorenson's sodium oxalate is 250 
milligrams of gold, and by titrating a solution of this amount of oxalate in 250 cc. 
of water, aciduated with a few drops of sulphuric acid, the oxidizing value of the 
permanganate solution is obtained in terms of gold. 

The value of the precipitating reagent and relative oxidizing value of the 
permanganate solution can be checked very accurately by adding a measured 
quantity of the reagent to an excess of gold chloride, filtering, washing thoroughly, 
incinerating and weighing the precipitate obtained in a tared porcelain crucible. 

Procedure. In carrying out the determination of a gold solution, a meas- 
ured or weighed portion is freed of oxidizing agents, a measured amount of the 
standard precipitating reagent added in slight excess of the amount required to 
decolorize the solution, and digestion on a steam bath or hot plate continued until 
the gold settles out, leaving a clear liquid. A few drops of sulphuric acid may be 
then added and, without filtering, titration performed. The gold value of the 
quantity of reagent added, minus that found of the excess of reagent, is the gold 
content of the amount of the sample taken. 

Iodide Method 

Small quantities of gold are determined by Gooch and Morley's iodide 
method.* A measured or weighed portion of the gold solution is treated, 
as has been described for removal of oxidizing agents, with an excess of free 
hydrochloric acid. Potassium iodide solution is run into the cold liquid until 
the gold precipitated as aurous iodide is completely dissolved. Starch solution is 
then added, and the amount of N/1000 thiosulphate required to decolorize the 
liquid noted. From this amount is deducted the amount of N/1000 iodine 
required to just produce a perceptible rose tint in the liquid. 

The reactions involved are AuCU-f 3KI =AuH-I,-f 3KC1 and l2-|-2Na2S20, = 
2NaI-|-Na,S40.. 

The gold value of the N/1000 solution of sodium thiosulphate should be deter- 
mined by performance of the operations of the method on a known quantity 
of gold, similar in amount and contained in a volume of solution approxi- 
mately equal to that of the analysis. 

Lenher's Method. By Lenher's method * of determining gold in solutions free of 
oxidizing agents, sulphurous acid of a reducing strength of 2-5 milligrams gold \xiT cc. is 
used as the rea^nt. The sulphurous acid requires frecjucnt standardizing by means 
of standard iodme or potassium iodide to which a definite amount of standard |)er- 
manganatc has been added or by a gold solution of known strength. Using starch 
as indicator, the iodine Hl)erated by addition of potassium iodide can be titrated by 
sulphurous acid. Bromine liberated by potassium bromide according to the equation, 
AuCl3-|-2KBr= AuCl-h2KCl-fBr2, can Ix; titrated by sulphurous acid. Excess of 
magnesium or sgdium chloride gives to auric chloride a yellow color which by sul- 
phurous acid can he titrated to the colorless or aurous state. These alkaline salts 
do not interfere in the potassium bromide or iodide reactions. 

lAmer. Jour. Sci., Oct., 1899, 261; Min. and Eng. World, 37, 853; Vol. Am., 
Sutton, 10th Ed.: " Assaying of Precious Metal," E. A. Smith. 
* Jour. Am. Cnem. Soc, June, 1913, 735. 



GOLD 197 



COLORIMETRIC METHODS 

Practical application of these methods is made in the estimation of gold in 
the liquors produced in the treatment of ores by the cyanide process. 

Prister's Method 

By Prister's method * a slight excess of copper solution is added to a 100 to 200- 
ec. portion of a cyanide solution in which the cyanide has been decomposed by 
boiling several minutes after acidifying with hydrochloric acid. Assurance of 
the presence of an excess of copper is made by spot test with a solution of potas- 
sium ferrocyanide. 

The copper solution is made by boiling for ten minutes in contact with copper 
shavings, a solution of 1 part blue vitriol and 2 parts salt in 10 parts of water, 
and adding a little acetic acid on cooling. A few drops of a 1 to 2 % sodiimi sul- 
phide solution are added, the liquid boiled for five minutes, the precipitate 
allowed to settle, and liquid separated by decantation on to a filter. The pre- 
cipitate in the beaker and on the filter is dissolved with 2J to 3 cc. of a 3 to 5% 
solution of potassium cyanide to which a few drops of potassium hydrate solu- 
tion has been added. 

Gold is precipitated from this cyanide solution (which may be turbid), by 
addition of 1 to 2 g ams of zinc dust and warming to 100° F. for half an hour. 
Liquid is separated by decantation through a filter. The residue on the filter 
and in the beaker is first treated with hydrochloric acid to dissolve zinc, then with 
10 cc. aqua regia, the reagent being passed several times through the filter. 
Stannous chloride solution is then added to the liquid diluted to 20 cc. Com- 
parison of the coloration produced is made with that from a standard solution of 
gold treated in the same manner. 

CassePs Method. By Cassel's method * 0.5 gram potassium bromate is mixed 
with 10 to 50 cc. of the cyanide solution and concentrated sulphuric acid added 
gradually with constant agitation until reaction commences. When the reaction 
stops, saturated solution of stannous chloride is added dropwise until the liquid is 
just colorless. The tint produced is compared with that from a standard gold 
solution treated in the same manner. 

Moir's Method. By Moir's method ' a measured quantity of the cyanide solution 
is oxidized by addition of 1 to 2 grams of sodium peroxide and boiling. If sufficient 
sodium peroxide is present, the brown spot produced by addition of a few drops of 
lead acetate will immediately dissolve. The lead-aluminum couple formed by addi- 
tion of aluminum j)owder precipitates gold which is filtered oflf. To the aqua regia 
solution of the precipitate a solution of stannous chloride is added drop by drop until 
the liquid is dissolved. The purple of Cassius tint developed is compared with per- 
manent standards composed of mixtures of solutions of copper sulphate and cobalt 
nitrate which have been adjusted to shades corresponding to those produced by known 
amounts of gold treated according to the method described. 

Bettel * filters suspended matter from the cyanide solution, adds a measured quan- 
tity of a strong solution of potassium cyanide which contains some cuprous cyanide and 
precipitates gold by the copper zinc couple produced by addition of a measured quantity 
of zinc fimie. The remainder of the method is the same as Prister's. 

1 Proc. Chem. Met. and Mm. Soc. of So. Af ., IV, 235, 1904. 

* Eng. and Mm. Journal. Oct. 31, 1903. 

» Proc. Chem. Met. and Min. Soc. of So. Af., Sept., 1913. 

* Min. World, 83, 102 and 35» 987. 



198 GOLD 

Dowsett's ^ factory test of barren cyanide solutions is capable of detecting 
variation in gold value of 1 cent per ton in solutions varying from one 
cent to about 15 cents per ton. To 500 cc. of the sample in a bottle with slight 
shoulder are added 10-15 cc. saturated sodium cyanide solution, 2 or 3 drops 
saturated lead nitrate solution and 1-2 grams 200-mesh fine zinc dust. The stop- 
pered bottle is shaken violently until the precipitate settles rapidly. Inverting 
the bottle allows the precipitate to settle into a casserole. Clear liquid is removed 
by decantation. Zinc is dissolved by hydrochloric acid added drop by drop until 
reaction ceases. A few drops excess hydrochloric acid and 3-5 drops dilute 
nitric acid (sp.gr. 1.18) are added and the liquid concentrated to 1-2 cc. The 
solution is transferred to a §-in. diameter test-tube, about 1 cc. of stannous chloride 
reagent added and grade of cyanide solution estimated by the tint obtained after 
one or two minutes standing. 1/1000 oz. gold per ton of original cyanide solution 
gives a very slight coloration; 15/10000 a slight yellow; 1/500 a slight pinkish 
yellow; 3/1000 a strong pink; 1/250 the purple of Cassius. Too much nitric acid 
hinders the production and the presence of mercury causes modification of the 
color. No more lead nitrate should be used than is sufficient to produce a rapidly 
settling precipitate. The stannous chloride reagent is a water solution contain- 
ing about 12i% crystals and 10% concentrated hydrochloric acid. 

PREPARATION OF PROOF GOLD 

Commercial gold may contain arsenic, antimony, selenium, tellurium, copper, 
lead, mercury, sUver, zinc, palladium, platinum and other metals of the platinum 
group. The method of making pure gold depends to a certain extent upon the 
character and quantity of impurities.* The method described assumes the 
raw material to be of extreme impurity. The metal is treated in 10-g am portions. 

When the metal contains silver its solution is effected most quickly by rolling 
extremely thin and annealing before treatment with acids. 

The strips, in a covered No. 6 casserole on a steam bath, are dissolved with a 
mixture of 5 cc. nitric and 50 cc. hydrochloric acid. If but little silver is present 
the quantity of hydrochloric acid may be decreased to 25 cc. The solution is 
evaporated to dryness and the casserole gently heated over a Bunsen flame until 
all the gold is reduced to metal. 

Digestion with ammonia will dissolve most of the silver and copper. After 
decanting the ammoniacal solution and washing with water, the gold is digested 
with hot nitric acid. If the solution is wine colored the digestion is continued 
for several hours, and reheated with fresh portions of acid until the absence of 
color indicates removal of palladium. The gold is now dissolved with 5 cc. of nitric 
and 15 to 20 cc. hydrochloric acids, evaporated to dryness, residue moistened with 
the least quantity of hydrochloric acid, dissolved with about 800 cc. water and 
liquid transferred to a 1000 cc. beaker. After the faint cloud of silver chloride 
settles to the bottom of the beaker, the clear liquid only is siphoned to another 
beaker, and allowed to stand another period of several days if it appears at all 
cloudy. The clear liquid is now siphoned into a 1000-cc. flask and sulphur dioxide 
gas passed until the gold is practically all precipitated. The gold is allowed to 

» Trans. I.M.M., 1912-13, 190; Met. and Chcm. Eng., July, 1914. 

«Eng. and Min. Jour., 68, 785, 1899; "Metallurgy of Gold/' Rose, 5th Ed.; Min. 
and Sci. Press, Nov. 14, 1903; ** Manual of Fire Assaymg," Fulton; " Assaying of Pre- 
cious Metals,'' Smith. 



GOLD 199 

settle, digested with hot nitric acid for a few minutes, washed by decantation 
several times, redissolved with aqua regia, solution transferred to a casserole, and 
nitric acid expelled by repeated evaporation to syiup with addition of hydro- 
chloric acid. The product of the second evaporation is moistened with the least 
quantity of hydrochloric acid, dissolved with water and solution transferred to 
a 1000-cc. be^er or Erlenmeyer flask. To the Uquid of about 500-cc. volume is 
added 11 grams of ammonium oxalate crystals. The beaker is permitted to 
remain on a steam bath until reaction is complete. The spongy mass of gold 
is now washed with hot water by decantation until free of salts. 

The gold is dried, melted in a clay crucible which has previously been thinly 
glazed with borax glass and poured out into a mold of charcoal, graphite and clay 
or iron polished with graphite. 

The ingot, which will have a volume of half a cubic centimeter, is cleaned by 
paring with a knife and rolled or hammered into a thin sheet. The rolls or 
hammer should be clean, bright and free of grease. 

The gold, cut into convenient strips, is digested for several hours with hydro- 
chloric acid and finally washed thoroughly with distilled water. 

The dried gold thus prepared may be considered 1000 fine. 



IODINE 

Wilfred W. Scott 

I, at.wt. 126.92; sp.gr. 4.94%^"^°; m.p. 113.5^;* b.p. 184.4* C; acids, HI, 

HIO, HIO„ HIO4. 

DETECTION 

The element may be recognized by its physical properties. It is a grayish 
black, crystalline solid, with metallic luster, brownish-red in thin layers. It 
vaporizes at ordinary temperatures with characteristic odor. Upon gently 
heating the element the vapor is evident, appearing a deep blue when unmixed 
with other gases, and violet when mixed with air. It colors the skin brown. 
Chemically it behaves very similarly to chlorine and bromine. 

Free iodine colors water yellow to black, carbon disulphide violet, ether 
or chloroform a reddish color, cold starch solution blue. 

Tannin interferes with the usual tests for iodine, unless ferric chloride is 
present. 

Iodide. The dry powder, heated with concentrated sulphuric acid, evolves 
violet fumes of iodine. Iodine is liberated from iodides by solutions of As*, 
Sb», Bis Cu", Fe'", Cr«, H,Fe(CN)6, HNO2, CI, Br, H,0,, ozone. 

Insoluble iodides may be transposed by treatment with H2S, the filtered 
solution being tested for the halogen. 

lodate. The acidulated solution is reduced by cold solution of S0», or 
K4Fe(CN)e, (acidulated with dilute H,S04), or by Cu^CU, H^sO,, FeS04, etc. 
An iodate in nitric acid may be detected by diluting the acid with water, adding 
starch solution, then hydrogen sulphide water, drop by drop, a blue zone forming 
in presence of the substance. 

ESTIMATION 

The element is found free in some mineral waters; combined as iodides 
and iodates in sea water; in ashes of sea plants; small quantities in a number 
of minerals, especially in Chili saltpeter as sodium iodate, hence in the mother 
liquor from the Chilian niter works from which iodine is principally produced. 
Sea-weed ash (drift kelp, Laminaria digitata and L. stenophylla) is an important 
source of iodine. 

Free iodine, potassium iodide, iodoform, are the principal commercial products. 

Preparation and Solution of the Sample 

In dissolving the substance it will be recalled that free iodine is soluble in 
alcohol, ether, chloroform, glycerole, benzole, carbon disulphide, solutions of 
soluble iodides. One hundred cc. of water at 11° C. is saturated with 0.0182 
gram iodine, at 55° with 0.092 gram. 

* Circular 35 (2d ed.) U. S. Bureau of Standards. 

200 



IODINE 201 

Iodides of silver, copper (cuprous), mercury (mercurous), and lead are 
insoluble, also Til, Pdls. Iodides of other metals are soluble; those of bismuth, 
tin, and antimony, require a little acid to hold them in solution. 

lodates of silver, bariiun, lead, mercury, bismuth, tin, iron, chromium 
require more than 500 parts of water at 15° C. to hold them in solution, 
lodates of copper, aliuninum, cobalt, nickel, manganese, zinc, calciiun, strontiiun, 
magnesixmi, sodium, and potassiiun are more soluble. One hundred cc. of 
cold water dissolves 0.00385 gram AglOs and 0.000035 gram Agl at ordinary 
temperatures. 

Free Iodine (Commercial Crystals). Iodine is best brought into solution 
in a strong solution of potassium iodide according to the procedure described 
for standardization of sodiiun thiosulphate under Volumetric Methods. The 
iodine is now best determined volumetrically by titration with standard thio- 
sulphate or arsenic. 

Iodine or Iodides in Water. The sample of water is evaporated to about 
one-fourth its volume and then made strongly alkaline with sodium carbonate. 
The precipitated calcium and magnesium carbonates are filtered off and washed. 
The filtrate containing the halogens is evaporated until the salts begin to crystallize 
out. The hot concentrated solution is poured into three volimies of absolute 
alcohol and the resulting solution again filtered. The residue is washed four 
or five times with 95% alcohol. All of the bromine and iodine pass into the 
solution, whereas a large part of chlorine as sodium chloride remains insoluble 
and is filtered off. About half a cc. of 50% potassium hydroxide is added and 
a greater part of the alcohol distilled off with a current of air. The residue 
is concentrated to crystallization and again poured into three times its volimie 
of absolute alcohol and filtered as above directed. This time only one or two 
drops of potassium solution is added and the procedure repeated several 
times. The final filtrate is freed from alcohol by evaporation, the solution 
taken to dryness and gently ignited, then taken up with a little water and filtered. 
Iodine is determined in the filtrate, preferably by the volumetric procedure III, 
decomposition with nitrous acid, described imder Volumetric Methods, p. 206. 

Organic Substances. If only an iodide is present, the Carius method is 
followed; in presence of other halogens, the **lime method" is preferred. Details 
of these methods are given in the chapter on Chlorine xmder Preparation and 
Solution of the Sample, p. 121. 

Silver iodide cannot be separated from the glass of the combustion-tube 
by solution with ammonium hydroxide as is the chloride or bromide of silver. 
The compound, together with the glass, is collected upon a filter paper, and 
washed with dilute nitric acid, followed by alcohol; then dried at 100® C. 
After removing most of the iodide and the glass, the filter is ignited in a weighed 
porcelain crucible, the main bulk of the material then added, the substance 
fused and weighed as Agl + glass. The mass is then covered with dilute sul- 
phuric acid and a piece of pure zinc added. After several hom« (preferably over 
night) the excess zinc is carefully removed and the iodine solution decanted 
from the glass and metallic silver, and the residue washed by decantation. The 
silver is now dissolved in hot dilute nitric acid, then filtered from the residue of 
glass through a small filter. The glass and filter are ignited and weighed. The 
difference between the two weighings is due to silver iodide. 

Minerals. Phosphates. The substance is decomposed by digestion with 
1 : 1 sulphuric acid in a fiask through which a current of air passes to sweep out 



202 IODINE 

the iodine vapor into a solution of potassium hydroxide, the sample being 
boiled until all the iodine vapors have been driven into the caustic. lodates 
are converted to iodides by reduction with sulphurous acid. 

With the iodine content below 0.02%, a 50 to 100-gram sample should be 
taken. 

SEPARATIONS 

Separation of Iodine from the Heavy Metals. The heavy metals are pre- 
cipitated as carbonates by boiUng with solutions of alkali carbonates, the soluble 
alkaU iodide being formed. 

Iodine is liberated from combination by nitrous acid. 

Silver iodide may be decomposed by warming with metallic zinc and sul- 
phuric acid. 

Separation of Iodine from Bromine or from Chlorine.^ Details of sepa- 
ration and estimation of the halides in presence of one another are given in 
the chapter on Chlorine. Advantage is taken of the action of nitrous acid on 
dilute solutions, free iodine being liberated, while bromides and chlorides are 
not acted upon. 

The solution containing the halogens is place in a large, round-bottom flask 
and diluted to about 700 cc. Through a two-holed stopjier a glass tube passes 
to the bottom of the flask; through this tube steam is conducted to assist the 
volatilization of iodine. A second short tube connected to the absorption appa- 
ratus conducts the evolved vapor from the flask into a 5% caustic soda solution 
containing an equal volume of hydrogen peroxide (about 50 cc. of each). The 
absorption system may be made by connecting two Erlenmeyer flasks in series, 
the inlet tubes dipping below the solutions in the flasks. It is advisable to cool 
the receivers with ice. 

Two to 3 cc. of dilute sulphuric acid (1 : 1) and 25 cc. of 10% sodium nitrite 
solution are added to the liquid containing the halogens, the apparatus is immedi- 
ately connected, and the contents of the large flask heated to boiling, conducting 
steam into it at the same time. The iodine vapor is gradually driven over into 
the cooled receiving flasks. 

When the solution in the large flask has become colorless it is boiled for half 
an hour longer. The steam is now shut off, the flask disconnected from the receiv- 
ing flasks and the heat turned off. The contents of the receiving flasks are com- 
bined with the washing from the connecting tubes and the solution heated to 
boiling to expel, completely, hydrogen peroxide. The cooled liquid is acidifled 
with a little sulphuric acid and the solution decolorized with a few drops of sul- 
phurous acid. Iodine is now precepitated as silver iodide by adding an excess 
of silver nitrate and a little nitric acid and lx)iling the mixture to coagulate the 
precipitate. Agl is determined as directed on page 203. 

Chlorine and bromine remain in the large flask in combined form and may be 
determined in this solution if desired. 

Notes. Reactions: 2KI 4-2KNO, -I-4H2S04 = Ij -I-2N0 + 4KHSO4 +2H .0. 

2NaOH + l2 = Nal-fNalO + H20andNaIO-|-H202 = H26+02+NaI. 

Consult Separations in the chapter on Chlorine, p. 123. 

* References: Method of Jannasrh, Zeit. ftir anorg. Chem., 1, p. 144 (1892). Tread- 
well and Hall, ** Analytical Chemistry." F. A. Clooch '* Methods in Chemical Analysis." 



IODINE 203 

Separation of Iodine from Chlorine and Bromine by Precipitation as 
Palladous Iodide. The solution contammg the halogens is acidified with hydro- 
chloric acid, and palladous chloride solution added to the complete precipitation 
of the iodide. The compound is allowed to settle in a warm place for twenty- 
four hours or more and then filtered and washed free of the other halogens. 
It may now be dried and weighed as palladous iodide, Pdiz, or ignited in a 
ciurent of hydrogen, then weighed as metallic palladium and the equivalent 
iodine calculated. See Gravimetric methods. 



QRAVIMETRIC METHODS 
Precipitation as Silver Iodide 

The procedure is practically the same as that described for determining 
chlorine. 

Silver nitrate solution is added to the iodide solution, slightly acidified with 
nitric acid. The precipitate is filtered into a weighed Gooch crucible, then 
washed, dried, gently ignited, and weighed as silver iodide. 

AglX 0.5406 =1 or X 0.7071 =KI. 

Note. If filter paper is used in place of a Gooch crucible, the precipitate is 
removed and the filter ignited separately. A few drops of nitric and hydrochloric 
acid are added, the acids expelled by heat and the residue weighed as AgCl. This, 
multipUed by 1.638 =AgI. The result is added to the weight of the silver iodide, 
which is ignited and weighed separately. 

Determination of Iodine as Palladous Iodide 

This method is applicable for the direct determination of iodine in iodides 
in presence of other halogens. 

The method of isolation of iodine as the palladous salt has been given imder 
Separations. The salt dried at 100° C. is weighed as Pdlj. 

PdI,X0.704=I. 

Pdli ignited in a current of hydrogen is changed to metallic palladimn. 

PdX2.379=I. 

VOLUMETRIC METHODS 
Determination of Hydriodic Acid — Soluble Iodides 

Free hydriodic acid cannot be determined by the usual alkalimetric methods 
for acids. The procedures for its estimation, free or combined as a soluble 
salt, depends upon the liberation of iodine and its titration with standard sodium 
thiosulphate, in neutral or slightly acid solution; or by means of standard arsen- 
ious acid, in presence of an excess of sodium bicarbonate in a neutral solution. 
The following equations represent the reactions that take place: 
I. Thiosulphate. 2NaS20,+l2=2NaI+Na,S406. 

n. Arsenite. Na,AsO,+l2+H,0=Na;^As04+2HI. 



204 IODINE 

The free acid formed in the second reaction is neutralized and the reversible 
reaction thus prevented: 

HI+NaHCO, =NaI+H,0+C02. 

The presence of a free alkali is not permissible, as the hydrox>d ion would 
react with iodine to form iodide, hypoiodite and finally iodate, hence sodium 
or potassium carbonates cannot be used. Alkali bicarbonates, however, do not 
react with iodine. 

Standard Solutions. Tenth Normal Sodium Thiosulphate. From the 
reaction above it is evident that 1 gram molecule of tliiosulphate is equivalent 
to 1 atom iodine = 1 atom hydrogen, hence a tenth normal solution is equal to one- 
tenth the molecular weight of the salt per Hter, e.g., 24.822 grams Na2Sj03 -51120; 
generally a slight excess is taken — 25 grams of the crj'stallized salt. It is 
advisable to make up 5 to 10 liters of the solution, taking 125 to 250 grams 
sodimn thiosulphate crystals and making up to volmne with distilled water, 
boiled free of carbon dioxide. The solution is allowed to stand a week to 
ten days, and then standardized against pure, resublimed iodine. 

About 0.5 gram of the purified iodine is placed in a weighing bottle con- 
taining a known amount of saturated potassium iodide solution (2 to 3 grams 
of KI free from KIO» dissolved in about \ cc. of HjO), the increased weight 
of the bottle, due to the iodine, being noted. The bottle and iodine are placed 
in a beaker containing about 200 cc. of 1% potassium iodide solution (1 gram 
KI per 200 cc), the stopper removed with a glass fork and the iodine titrated 
with the thiosulphate to be standardized. 

Calculation. The weight of the iodine taken, divided by the cc. thio- 
sulphate required, gives the value of 1 cc. of the reagent; this result divided 
by 0.012692 gives the normality factor. 

Note. The thiosulphate solution may be standardized against iodine, which 
has been liberated from potassium iodide in presence of hydrochloric acid by a known 
amount of standard potassium bi-iodatc, a salt which may be obtained exceedingly 
pure. 

KIO,-HIOa-hl0KI-hllHCl = llKCl-h6H2O-h6l2. , 

A tenth normal solution contains 3.2496 grams of the pure salt per liter. (One 
cc. of this will liberate 0.012692 gram of iodine from potassium iodide.) The purity 
of the salt should be established by standardizing against thiosulphate, which has been 
freshly tested against pure resubUmed iodine. 

About 5 grams of potassium iodide (free from iodate) are dissolved in the least 
amount of water that is necessary to effect solution, and 10 cc. of dilute hydrochloric 
acid (1:2) are added, and then 50 cc. of the standard bi-iodate solution. The solu- 
tion is diluted to about 250 cc. and the liberated iodine titrated with the thiosulphate 
reagent; 50 cc. will be required if the reagents are exactly tenth normal. 

Tenth Normal Arsenite. From the second reaction above it is evident that 
AsiOj is equivalent to 2I2, e.g., to 4H, hence \ the gram molecular weight of 
arsenious oxide per liter will give a nonnal solution: 198^4 =49.5. 

4.95 grants of pure arsenious oxide is dissolved in a Uttle 20% sodium 
hydroxide solution, the excess of the alkali is neutralized with dilute sulphuric 
acid, using phenolphthalein indicator, the solution lx»ing just decolorized. Five 
hundred cc. of distilled water containing about 25 grams of sodium bicarbonate 
are added. If a pink color develops, this is destroyed with a few drops of 
weak sulphuric acid. The solution is now made to volume, 1000 cc. The 



IODINE 205 

reagent is standardized against a measured amount of pure iodine. The oxide 
may be dissolved directly in sodium bicarbonate solution. 

Note. Commercial arsenious oxide is purified by dissolving in hot hydrochloric 
acid, filtering the hot saturated solution, cooling, decanting off the mother liquor, 
washing the deposited oxide with water, drying and finally suoliming. 

Starch Solution. Five grams of soluble starch are dissolved in cold water, 
the solution poured into 2 liters of hot water and boiled for a few minutes. 
The reagent is kept in a glass-stoppered bottle. 

Iodides are decomposed and iodine determined by one of the following 
procedures : 

I. Decomposition of the Iodide by Ferric Salts 

The method takes advantage of the following reaction: 

Fe,(S04)3+2KI =K2S04+L+2FeS04. 

The procedure enables a separation from bromides, as these are not acted 
upon by ferric salts. 

Procedure. To the iodide in a distillation flask is added an excess of ferric 
ammonimn alum, the solution acidified with sulphuric acid, then heated to 
boiling, and the iodine distilled into a solution of potassium iodide. The free 
iodine in the distillate is titrated with standard thiosulphate, or by arsenious 
acid in presence of an excess of sodium bicarbonate. 

The reagent is added from a burette until the titrated solution becomes a 
pale yellow color. About 5 cc. of starch solution are now added and the titra- 
tion continued until the blue color of the starch fades and the solution becomes 
colorless. 

One cc. of tenth normal reagent =0.012692 gram iodine, equivalent to 
0.012793 gram HI, or 0.016602 gram KI. 

II. Decomposition with Potassium lodate ^ 

The reaction with potassium iodate is as follows: 

5KI+KI0,+6HC1 =KC1+3H,0+3I,. 

It is evident that i of the titration for iodine would be equal to the iodine 
of the iodide, hence 1 cc. of tenth normal thiosulphate is equivalent to 0.012692 
Xt =0.01058 gram iodine due to the iodide. The procedure is as follows: 

Procedure. A known amount of tenth normal potassium iodate is added 
to the iodide solution, in sufficient amount to liberate all of the iodine, com- 
bined as iodide, and several cc. in excess. Hydrochloric acid and a piece of calcite 
are added. The mixture is boiled until all of the liberated iodine has been 
expelled. To the cooled solution 2 or 3 grams of potassiiun iodide are added 
and the liberated iodine, corresponding to the excess of iodate in the solution, 
is titrated with standard thiosulphate. If 1 cc. of thiosulphate is equal to 1 
cc. of the iodate, then the total cc. of the iodate used, minus the cc. thio- 

^H. Dietz and B. M. Margosches, Chem. Ztg., 2, 1191, 1904. Treadwell 
and Hall, "Analytical Chemistry," Vol. 2. 



206 



IODINE 



sulphate required in the titration gives a difference due to the volume of iodate 
required to react with the iodide of the sample. 

One cc. of N/10 KIO, =0.01058 gram I in KI. 
Note. Tenth normal potassium iodate contains 3.5675 grams KIOs per 1000 cc. 



III. Decomposition of the Iodide with Nitrous Acid (Fresenius)^ 

Nitrous acid reacts with an iodide as follows: 

2HNO:+2HI =2NO+2H,0+I,. 

Since neither hydrochloric nor hydrobromic acids are attacked by nitrous 
acid, the method is applicable to determining iodine in presence of chlorine and 

bromine; hence is useful for determining small amounts of iodine in 
mineral waters containing comparatively large amounts of the other 
halogens. 

Nitrous Acid. The reagent is prepared by passing the gas 
into strong sulphuric acid until saturated. 

Procedure. The neutral or slightly alkaline solution of the 
iodide is placed in a glass-stoppered separatory funnel, Fig. 39, 
and slightly acidified with dilute sulphuric acid. A little freshly 
distilled colorless carbon disulphide (or chloroform) is added, then 
10 drops of nitrous acid reagent. The mixture is well shaken, the 
disulphide allowed to settle, drawn off from the supernatant solu- 
tion and saved for analysis. The liquor in the funnel is again 
extracted with a fresh portion of disulphide and if it becomes dis- 
colored it is drawn off and added to the first extract. If the 
extracted aqueous solution appears yellow, it must be <igain treated 
with additional carbon disulphide until all the iodine has been 
removed (e.g., until additional CSj is no longer colored when shaken 
with the solution). The combined extracts are washed with three 
or four portions of water, then transferred to the filter and again 
washed until free from acid. A hole is made in the filter and the disulphide 
allowed to run into a small beaker and the filter washed down with about 5 cc. 
of water. Three cc. of 5% sodium bicarbonate are added and the iodine titrated 
with N/20 or N/50 standard thiosulphate, the reagent being added until the 
reddish-violet carbon disulphide becomes colorless. 

The sodium thiosulphate used is standardized against a known amount of 
pure potassium iodide treated in the manner described above. 

One cc. N/20 Na2S20,=. 00635 gram I, 1 cc. N/50 Na,SA=. 002538 gram I. 




Fig. 39. 



IODINE 207 

IV. Liberation of Iodine by Means of Hydrogen Peroxide and 

Phosphoric Acid ^ 

Principle. Iodine is liberated from an iodide by addition of hydrogen per- 
oxide to the solution acidified with phosphoric acid, the iodine distilled into 
potassium iodide and titrated with thiosulphate. 

Procedure. Fifty cc. of the iodide solution are mixed with 5 cc. of pure 
phosphoric acid and 10 to 20 cc. hydrogen peroxide added, the mixture being 
placed in a round-bottomed flask, connected with a short condenser, delivering 
into two absorption vessels containing a 10% solution of potassiimi iodide. 
A current of air is drawn through the apparatus, and the contents of the flask 
gradually heated to boiling. The iodine is absorbed in the potassium iodide 
solution and titrated as usual with standard sodium thiosulphate. Twenty 
minutes' heating is generally sufficient. 

One cc. NaiSjOa =0.012692 gram I, or 0.016602 gram KI. 

Note. Iodine in urine m&y be determined by evaporating to 1/10 its volume. 
After adding an excess of sodium hydroxide^ the mixture is taken to dryness and 
gently ignited. The ash may be used for the iodine determination. 

V. Oxidation of Combined Iodine with Chlorine. (Mohr's 

. Modification of Duprfe's Method)2 

When a solution of potassiiun iodide ia treated with successive amounts 
of chlorine water, iodine is liberated, which reacts with an excess of chlorine 
with formation of chloride of iodine (ICl) and with greater excess the penta- 
chloride (ICU) which is changed in presence of water to iodic acid (HlOg). 

Procedure. The weighed iodide compound is brought into a stoppered 
flask, and chlorine water delivered from a large burette until all yellow color has 
disappeared. A drop of the mixture brought in contact with a drop of starch 
solution should produce no blue color. Sodium bicarbonate is now added until 
the mixture is slightly alkaline, followed by an excess of potassiiun iodide and 
4 to 5 cc. of starch reagent. Standard thiosulphate is now added until the 
blue color is removed. The excess of chlorine water is thus ascertained. From 
the value of the chlorine reagent the iodine of the sample may readily be 
calculated. 

The chlorine water is standardized by running 25 to 50 cc. of the reagent 
into potassium iodide solution (see procedure for bromides, p. 81), and 
titrating the liberated iodine with standard sodium thiosulphate. The value 
of the reagent in terms of thiosulphate are thus ascertained and from this the 
value per cc. in terms of iodine. 

OTHER METHODS 

Volhard's Method for Determining Iodides 

This procedure is very similar to those for determining chlorine or bromine, 
with the exception that silver iodide formed will occlude both the iodide solu- 

»E. Winterstein and E. Herzfeld, Zeit. Physiol. Chem., 68, 49-51, 1909. Chem. 
Zentralbl., (1), 473-474, 1910. 

* Sutton, "Volumetric Analysis,", 10th Ed. 



208 IODINE 

tion and silver nitrate unless the additions of the silver salt are made in small 
portions with vigorous shaking. 

Standard silver nitrate is added to the solution in a glass-stoppered flask, 
shaking vigorously with each addition. As long as the solution appears milky 
the precipitation is incomplete. When the silver iodide is coagulated and the 
supernatant liquid appears colorless, ferric alum solution is added, and the 
excess of silver nitrate titrated with potassium sulphocyanate until the char- 
acteristic reddish end-point is obtained. 

The iodine is calculated from the amount of silver nitrate required. E.g., 
total AgNOa added, minus excess determined by KCNS=cc. AgNOa required 
by the iodine. 

Note. The ferric salt oxidizes hydriodic acid with separation of iodine, whereas 
the silver iodide is not acted upon, hence the indicator is added after all the iodide 
has combined with silver. 

VI. Determination of lodates 

The procedure is the reciprocal of the one for determination of iodide by 
means of an iodate : 

Reaction. KIO3+5KH-6HCI -eKCl-hSHaO-j-SL. 

Procedure. The solution containing the iodate is allowed to run into an 
excess of potassium iodide solution containing hydrochloric acid. The liber- 
ated iodine is titrated with sodium thiosulphate as usual. 

One cc. N/10 Naj&Oa =0.002932 gram HlOa, or 0.003567 gram KIO3. 

VII. Determination of Periodates 

The procedure is the same as that described for iodates, the reaction in this 
case, however, being as follows: 

KIO4+7KH-8HCI =8KC1+4H20-|-4I,. 

From the equation it is evident that 1 gram molecule of the iodate is equiv- 
alent to 8 atoms of iodine =8 atoms of hydrogen, hence J the molecular weight 
per liter of solution would equal a normal solution. Therefore, 1 cc. of a tenth 
normal solution would contain 0.019193^8 =0.002399 gram HIO4. 

One cc. N/10 NajSaO, =0.002399 gram HIO4, or =0.002849 gram HI04-2H20, 
or =0.002875 gram KIO4. 

VIII. Determination of lodates and Periodates in a Mixture 

of the Two 

The procedure depends upon the fact that an iodate does not react with 
potassium iodide in neutral or slightly alkaline solutions, whereas a pcriodate 
undergoes the following reactions : 

KI04+2KI-hH30 =2KOH-|-KI03+l2. 

Procedure. The sample, dissolved in water, is divided into two equal 
portions. 

A, To one portion a drop of phenolphthalein indicator is added and the 



IODINE 209 

solution made just faintly alkaline by addition of alkali to acid solutions oi' 
hydrochloric acid to alkaline solution, as the case may require. Ten cc. of 
cold saturated solution of sodium bicarbonate are added and an excess of potas- 
sium iodide. The liberated iodine is titrated with tenth normal arsenious acid.* 
(Na2S20s will not do in this case, as the solution is alkaline.) 

One cc. N/10 As,0, =0.0115 gram KIO4. 

B. To the other portion potassiimi iodide is added in excess and the solu- 
tion made distinctly acid. The liberated iodine is titrated with standard sodium 
thiosulphate. (AsjOj will not do.) 

Calculation. In the acid solution, B, both iodates and periodates are 
titrated, whereas in the alkaline solution, A^ only the periodates are affected. 
From the reactions in VII and VIII it is evident that 1 cc. NslS^Oi =4 cc. AsiOa 
for the periodate titration, hence 

Cc. NajSsO*— cc. AsjOjX4=cc. thiosulphate due to KIOi. 

The difference, multiplied by 0.003567 = grams KIOi in the sample. 

^ In alkaline solutions the arsenious acid titration must be made, whereas in acid 
solutions potassium thiosulphate is used. 



IRON 

Wilfred W. Scott 

Fe, at.wt. 55.84; sp.gr. 7.85-7.88; m.p. pure 1530%* wrought 1600°,« white 
pig 1075°^' gray pig 1275%^ steel 1375*»; ^ fe.p. »450*» ^ C; oxides FeO, 
FesOs, FcaO*. 

DETECTION 

Ferric Iron. The yellow to red color in rocks, minerals, and soils is gen- 
erally due to the presence of iron. 

Hydrochloric acid solutions of iron as ferric chloride are colored yellow. 

Potassium or ammonium sulphocyanate produces a red color with solutions 
containing ferric iron. Nitric acid and chloric acid also produce a red color with 
potassium or ammonium sulphocyanate. This color, however, is destroyed by 
heat, which is not the case with the iron compound. The red color of ferric 
iron with the cyanate is destroyed by mercuric chloride and by phosphates, 
borates, certain organic acids, and their salts, e.g., acetic, oxalic, tartaric, citric, 
racemic, malic, succinic, etc. 

Potassium ferrocyanide, K4Fe(CN)6, produces a deep blue color with ferric 
salts. 

Salicylic acid added to the solution of a ferric salt containing no free mineral 
acid gives a violet color. Useful for detecting iron in alum and similar products. 

Ferrous Iron. Potassium Ferricyanide, K3Fe(CN)6, gives a blue color 
with solutions of ferrous salts. 

Distinction between Ferrous and Ferric Salts. 

KCNS gives red color with Fe'" and no color with Fe". 

Bl3Fe(CN)e gives a blue color with Fe" and a brown or green with Fe"'. 

NH4OH, NaOH or KOH precipitates red, Fe(OH), with Fe"' and white, 
Fe(0H)2 with Fe" turning green in presence of air due to oxidation.* 

Sodium peroxide produces a reddish-brown precipitate of Fe(0H)3 with 
either ferrous or ferric salt solutions, the former being oxidized to the higher 
valence by the peroxide. Chromium and aluminum remain in solution, if present 
in the sample. 

ESTIMATION 

Iron is so widely diffused in nature that its determination is necessary' in 
practically all complete analyses of ores, rocks, minerals, etc. It is especially 
important in the evaluation of iron ores for the manufacture of iron and steel. 
Among the ores of iron the following are more common: 

Oxides. Red hematite, FejOa,* brown hematite, 2Fe203 -31120; black mag- 
netite or magnetic iron ore, Fea04. Ferric oxide with varying amounts of water 

» Circular 35 (2d Ed.) U. S. Bureau of Standards. 

• D. Van Nostrand's Chemical Annual. — Dlsen. 

• The green salt id a hydrate of FejOi. The white precipitate can be obtained in 
absence of air or by using HjSOs to take up oxygen in solution. 

210 



IRON 211 

forms the substances known as hematite, gothite, liraonite, yellow ochre, bog 
iron ore. 

Sulphide. Iron pyrites or "fool's gold," FeS^; pyrrhotite, FeS. 

Carbonates. Spatic iron ore, FeCOa; combined with clay in clay ironstone 
with bituminous material as "black band." 

Iron is determined in the cinders and in iron ore briquettes from burned iron 
pyrites, by-products of sulphuric acid. 

It is found as an impurity in a large number of commercial salts and in the 
mineral acids. 

Preparation and Solution of the Sample 

The material should be carefully sampled and quartered down according 
to the general procedure for sampling. Ores should be ground to pass an 80- 
mesh sieve. In analysis of metals, both the coarse and fine drillings are taken. 

The following facts regarding solubility should be remembered: The element 
is soluble in hydrochloric acid and in dilute sulphuric acid, forming ferrous 
salts with liberation of hydrogen. It is insoluble in concentrated, cold sulphuric 
acid, but is attacked by the hot acid, forming ferric sulphate with Uberation of 
SO2. Moderately dilute, hot nitric acid forms ferric nitrate and nitrous oxide; 
the cold acid gives ferrous nitrate and ammonium nitrate or nitrous oxide or 
hydrogen. Cold, concentrated nitric acid forms "passive iron," which remains 
insoluble in the acid. The oxides of iron are readily soluble in hydrochloric acid, 
if not too strongly ignited, but upon strong ignition the higher oxides dissolve 
with extreme difficulty. They are readily soluble, however, by fusion with 
acid potassium sulphate followed by an acid extraction. Silicates are best 
dissolved by hot hydrochloric acid containing a few drops of hydrofluoric acid 
or by fusion with sodium and potassium carbonates, followed by hot hydro- 
chloric acid. 

Soluble Iron Salts. Water solutions are acidified with HCl or H2SO4, so 
as to contain about 3% of free acid. 

Ores. The samples should be pulverized to pass an 80- to 100-mesh sieve. 

Sulphides, Ores Containing Organic Matter. One- to 5-gram samples 
should be roasted in a jjorcelain crucible over a Bunsen flame for about half 
an hour, until oxidized. The oxide is now dissolved as directed in the following 
procedure. 

Oxides, Including Red and Brown Hematites, Magnetic Iron Ore, Spatose 
Iron Ore, Roasted Pyrites, and Iron Ore Briquettes. One to 5 grams 
of the ore, placed in a 400-cc. beaker, is dissolved by adding twenty times 
its weight of strong hydrochloric acid with a few drops of 5% stannous chloride 
solution. Addition of 4 or 5 drops of HF is advantageous if small amounts 
of silica are present. The solution is covered with a watch-glass and heated to 
80 or 90° C. until solution is complete. Addition of more stannous chloride 
may be necessary, as this greatly assists solution. An excess sufficient to com- 
pletely decolorize the solution necessitates reoxidation with hydrogen peroxide, 
hence should be avoided. If a colored residue remains, it should be filtered 
off, ignited and fused with a mixture of NaaCOs and KjCOa in a platinum cru- 
cible. The fusion dissolved in dilute HCl is added to the main filtrate. 

Note. The ore placed in a porcelain boat in a red-hot combustion tube may be 
reduced with hydrogen (taking precaution first to sweep out oxygen with COj) and 
after coolins in an atmosphere of hydrogen the reduced iron may 1^ dissolved in acid 
and titratedf. 



212 IRON 

Iron Silicates. One to 5 grams of the material, placed in a deep plati- 
num crucible, is treated with ten times its weight of 60% HF and 3 to 4 drops 
of cone. H2SO4. The mixture is evaporated to near dryness on the steam bath 
and taken up with dilute sulphuric acid or hydrochloric acid. The latter acid 
la the best solvent for iron. ^ 

Fusion with Potassium Bisulphate. The sample is mixed with ten times 
its weight of the powdered bisulphate and 2-3 cc. of concentrated sulphuric acid 
added. A porcelain or silica dish will do for this fusion. The fusion should 
be made over a moderate flame and cooled as soon as the molten liquid becomes 
clear. Complete expulsion of SOj should be avoided. It may be necessary 
to cool and add more cone, sulphuric acid to effect solution. Iron and aliunina 
completely dissolve, but silica remains imdissolved. The melt is best cooled by 
pouring it on a large platinum lid. 

Fusioa with Carbonates of Sodium and Potassium. The residues insoluble 
in hydrochloric acid are fused with 5 parts by weight of the fusion mixture 
(NaiCOa+KjCOa) in a platinum crucible. The M6ker blast will be necessary. 
When the effervescence has ceased and the melt has become clear, the crucible 
is removed from the flame, a platinum wire inserted and the melt cooled. Upon 
gently reheating, the fuse may be readily removed by the wire in a convenient 
form for solution in dilute hydrochloric acid. 

The bisulphate fusion is recommended for fusion of residues high in iron 
and alumina. It is an excellent solvent for ignited oxides of these elements. 
The carbonate fusions are adapted to residues containing an appreciable amount 
of silica. 

Iron and steel are best dissolved in hydrochloric acid with a few drops of 
nitric acid. The iron hydroxide should be precipitated or the solution taken to 
dryness to expel the nitric acid followed by resolution in dilute hydrochloric 
acid or sulphuric acid. 

The finer the material the more rapid its solution is a fact that should be remem- 
bered in all cases. 



IRON 213 

QRAVIMETRIC METHODS FOR THE DETERMINATION 

OF IRON 

The gravimetric determination of iron may be made from solutions practi- 
cally free from other metals. A number of elements such as phosphorus, 
arsenic, molybdenum, tungsten, vanadium, and the like, form fairly stable 
comjjounds with iron in neutral or slightly alkaline solutions, whereas others, 
such as lead, copper, nickel, cobalt, sodium, and potassium may be occluded 
in the ferric hydrate precipitate and are removed only with considerable diffi- 
culty. Alimiinum, chromium, and several of the rare earths are precipitated 
with iron, if present. These facts taken into consideration, the volumetric 
methods are generally preferred as being more rapid and trustworthy. 

Determination of Iron as Fe203 

Iron is precipitated as the hydroxide and ignited to the oxide, FejOj, in 
which form it is weighed. 

Reactions. FeCl,+3NH40H =Fe(OH),+3NH4Cl. 

2Fe(0H),+heat =Fe203+3H20. 

Procedure. One-gram sample or a larger amount of material if the iron 
content is low, is brought into solution with hydrochloric acid, aqua regia, or by 
fusion with potassium carbonate or potassium acid sulphate, as the case may 
require. Silica is filtered off and the acid solution treated with H2S if members 
of that group are present. The filtrate is boiled to expel H2S and the iron 
oxidized to ferric condition by boiling with 5 cc. concentrated nitric acid. 

Absence of Aluminum and Chromium. About 1 gram of ammonium 
chloride salt or its equivalent in solution is added, the volume made to about 
200 cc. and ammonium hydroxide added in slight excess to precipitate Fe(0H)3. 
The solution is boiled for about five minutes, then filtered through an ashless 
filter. (S. & S. 589 is good for this purjjose.) 

If Aluminum and Chromium are Present. In place of ammonium hydroxide 
powdered sodium peroxide is added in small portions until the precipitate first 
formed clears, the solution being cold and nearly neutral. It is diluted to about 
300 cc. and boiled ten to fifteen minutes to precipitate the iron. Aluminum and 
chromium are in solution. (Mn will precipitate with Fe, if present.) The 
precipitate is filtered onto a rapid filter and washed with hot water. 

Second Precipitation. In either case dissolve the precipitate with the 
least amount of hot dilute hydrochloric acid and wash the paper free of iron. 
Add a few cc. of 10% ammonium chloride solution and reprecipitate the hydroxide 
of iron by adding an excess of ammonium hydroxide, the volume of the solu- 
tion being about 200 cc. Washing the precipitate by decantation is advisable. 
Three such washings, 100-cc. portions, followed by two or three on the filter 
paper, will remove all impurities. 

Ignition. The precipitate is ignited wet over a low fiame, gradually in- 
creasing the heat. Blasting is not recommended, as the magnetic oxide of 
iron, Fe804, wUl form with high heating. The oxide heated gently appears a 
reddish-brown. Higher heat gives the black oxide, Fe804. Twenty minutes' 
ignition, at red heat, is^sufficient. 



214 IRON 

The crucible, cooled in a desiccator, is weighed and FcjOs obtained. 

Factors. Fe,0, X 0.6994 = Fe. 
Fe^O, X 0.8998 =FeO. 

Precipitation of Iron with '' Cupferron," Amino nitrosophenyl- 

hydroxy lamine ^ 

By this procedure iron may be precipitated directly in acid solution in 
presence of a number of elements. Mercury, lead, bismuth, tin, and silver 
may be partially precipitated. Copper precipitates with iron, but may be 
easily removed by dissolving it out with ammonia. The method is especially 
adapted for separation of iron from alimiinum, nickel, cobalt, chromium, cadmium, 
and zinc. 

Procedure. The solution containing the iron is made up to 100 cc. and 20 
cc. of concentrated hydrochloric acid added. To this cool solution (room 
temperature) Baudisch's reagent, cupfcrron, is slowly added with constant 
stirring, until no further precipitation of iron takes place, and crystals of the 
reagent appear. The iron precipitate is a reddish-brown. Copper gives a 
grayish-white flocculent compound. An excess of the reagent equal to one-fifth 
of the volume of the solution is now added, the precipitate allowed to settle 
for about fifteen minutes, then poured into a filter paper and washed, first with 
2N. HCl, followed by water, then with ammonia and finally with water. The 
drained precipitate is slowly ignited in a porcelain or platiniun crucible and the 
residue weighed as FcjOa 

FeaOa X 0.6994 =Fe. 

Notes. Baudisch's reagent, amino nitrosophenyl-hydroxylamine (cupferron), is 
made by dissolving 6 grams of thi salt in water and diluting to 100 cc. The reagent 
keeps for a week if protected from the light. It decomposes in the light, forming 
nitrobenzine. Turbid solutions should be filtered. 

The precipitates of copper or iron are but slowly attacked by twice normal hydro- 
chloric acid in the cold, but decomposed by hot acid, hence the solution and reagent 
should be cold. 

Cold, dilute potassium carbonate solution, or ammonium hydroxide, have no action 
on the iron precipitate; the copper compound dissolves readily in ammonia. Alkaline 
hydroxide causes rapid decomposition. 

The precipitation is best made in comparatively strong acid solutions (HCl, H2SO4, 
or acetic acid). 

VOLUMETRIC DETERMINATION OF IRON IN ORES AND 

METALLURGICAL PRODUCTS 

General Considerations. Two general procedures are commonly employed 
in the determination of iron. 

A . Oxidation of ferrous to ferric condition by standard oxidizing agents. 

B. Reduction of ferric iron to ferrous condition. 

The sample is dissolved as directed under Preparation and Solution of the 
Sample. 

10. Baudisch, Chem. Ztg., 33, 1298, 1905. Ibid., 36, 913, 1911. O. Baudisch 
and V. L. King, Jour. Ind. Eng. Chem., 3, 027, 1911. 



IRON 215 



Determination of Iron by Oxidation Methods 

Some modification of either the dichromate or permanganate methods is 
commonly employed in the determination of iron by oxidation. To accomplish 
this quantitatively, the iron must be reduced to its ferrous condition. This 
may be accomplished in the following ways: 

1. Reduction by Hydrogen Sulphide. During the course of a complete anal- 
ysis of an ore, H,S is passed into the acid solution to precipitate the members of that 
group (Hg, Pb, Bi, Cu, Cd, As, Sb, Sn, Pt, Au, Se, etc.). The filtrate contains iron 
in the reduced condition suitable for titration with either dichromate or per- 
manganate, the excess of HjS having been boiled off. If the expulsion of HjS 
is conducted in an Erlenmeyer flask there is little chance for reoxidation of the 
iron during the boiling. Reduction by H2S is very effective and is frequently 
advisable. This is the case when titanium is present, since this is not reduced 
by H2S, but by methods given below. Arsenic, antimony, copper, and platinum, 
which, if present would interfere, are removed by this treatment. 

Reaction. 2FeCl,+H,S =2FeCl2+2HCl+S. 

2. Reduction with Stannous Chloride. SnCU solution acts readily in 
a hydrochloric acid solution of the ore; the reduction of the iron is easily noted 
by the disappearance of the yellow color. The excess of the reagent is oxidized 
to SnCU by addition of HgCU. 

Reactions. 1. 2FeCl3+SnCl,=2FeCl,+SnCl4. 

2. Excess SnCl2+2HgCU=SnCl4+2HgCl precipitated. 

An excess of SnCU is advisable, but a large excess is to be avoided, as a 
secondary reaction would take place, as follows: 2SnCl2+2HgCl2=2SnCl4-|-2Hg. 
This reaction is indicated by the darkening of the solution upon the addition of 
HgCU. Precipitation of metallic mercury would vitiate results. The solution 
should be cooled before addition of mercuric chloride. About 15-20 cc. of sat- 
urated mercuric chloride, HgCU, solution should be sufficient. 

3. Reduction by a Metal such as Test Lead, Zinc, Magnesium, Cadmium, 
or Aluminum, in Presence of Either Hydrochloric Acid or Sulphuric Acid. 
The former acid is preferred with the dichromate titration, and the latter with 
the permanganate. Two methods of metallic reduction are in common use — 
reduction by means of test lead, and reduction with amalgamated zinc by means 
of the Jones reductor.- 

(a) Reduction with Test Lead. By this method copper is precipitated 
from solution and small amounts of arsenic and antimony expelled. Sufficient 
test lead is added to the acid ferric solution to completely cover the bottom of 
the beaker. The solution is covered and boiled vigorously until the yellow color 
has completely disappeared, and the solution is colorless. The reduced iron 
solution, cooled, is decanted into a 600-cc. beaker, the remaining iron washed 
out from the lead mat by several decantations with water; two or three 50-cc. 
portions of water should be sufficient; the washings are added to the first portion. 
If the solution becomes sUghtly colored, a few drops of stannous chloride, SnCU, 
solution are added, followed by 10 cc. mercuric chloride, HgCh, solution. The 
sample is now ready for titration. 



216 IRON 

(6) Reduction with Zinc, Using the Jones Redactor. The acid solution of 
iron, preferably sulphuric acid, is passed through a column of amalgamated zinc* 
The hydrogen evolved in presence of the zinc reduces the ferric iron to ferrous 
condition. The procedure is described in detail under the Permanganate Method 
for Determination of Iron, page 218. Titanium if present will also be reduced. 

4. Reduction with Sulphurous Acid, Sodium Sulphite or Metabisulphite. 
SO2 gas is passed into a neutral solution of iron, since iron is not reduced 
readily in an acid solution by this method. The excess SO2 is expelled by acidi- 
fying the solution and boiling. 

6. Reduction with potassium iodide, the liberated iodine being expelled by 
heat. 

In the solution of the ore with stannous chloride and hydrochloric acid, if an 
excess of the former has been accidentally added, it wiU he necessary to oxidize the 
iron before reduction. This may he accomplished hy addition of hydrogen peroxide 
until the yellow color of ferric chloride appears {or hy addition o/KMn04 solution), 
the excess H2O2 may he removed by boiling. The iron may now he reduced hy one of 
the above methods. 

Volumetric Determination of Iron by Oxidation with 

Potassium Dichromate 

Principle. This method depends upK)n the quantitative oxidation of ferrous 
salts in cold acid solution (HCl or H2SO4) to ferric condition by potassium 
dichromate, the following reaction taking place : 

6FeCl2+K2Cr207+14HCl=6FeCl,+2CrCU+2KCl+7H20. 

Potassium ferricyanide is used as an outside indicator. This reagent pro- 
duces a blue compound with ferrous salts and a yellowish-brown with ferric. 
The chromic salt formed by the reaction with iron colors the solution green. 

Reagents Required. Standard Potassium Dichromate. When oxygen 
reacts with ferrous salts, the following reaction takes place : 

6FeCl2+6HCl-F30 =6FeCU-F3H20. 

Comparing this reaction with that of dichromate, it is evident that a normal 
solution of dichromate contains one-sixth of the molecular weight of K2Cr207 
per liter, namely, 49.033 grams. For general use it is convenient to have two 
strengths of this solution, N/5 for ores high in iron and N/10 for products con- 
taining smaller amounts. 

Standardization. For N/5 solution 9.807 grains of the recrystallized de- 
hydrated salt are dissolved and made up to x)ne liter; N/10 potassium dichro- 
mate contains 4.903 grams of the pure salt per liter. It is advisable to allow 
the solution to stand a few hours before standardization. The Sibley iron 
ore furnished by the U. S. Bureau of Standards, Washington, D. C, is recom- 
mended as the ultimate standard. Other ores uniform in iron may be standardized 
against the Sibley ore and used as standards. The ore in cjuestion contains 
69.20% Fe (1914). For accurate work it is desirable to use a chamber burette 

* Amalgamated zinc is best prepared by dissolving 5 grams of mercury in 25 cc. 
of concentrated nitric acid with an equal volume of water, 250 cc. of water are added 
and the solution poured into 500 grams of shot zinc, 'iO-mesh. When thoroughly 
amalgamated the solution is poured off, and the zinc dried. 



IRON 217 

with graduations from 75 to 90 cc. in tenths and from 90 to 100 in twentieths of 
a CO. A titration of 90 to 100 cc. of the dichromate would require 0.9 to 1.1 
gram of iron for a fifth normal solution and half this amount for a tenth normal 
solution of dichromate. In the first case 1 .4 gram of Sibley iron ore should 
be taken and for N/10 0.7 gram of the ore. The ore is best dissolved in strong 
HCl, adding a few drops of stannous chloride solution and heating just below 
boiling. In case of an ore or iron ore briquette, containing silica in an appre- 
ciable amount, a carbonate fusion of the residue may be necessary. Reduction 
and titration of the ore is done exactly as prescribed under Procedure below. 

The equivalent iron in the ore divided by the cc. titration required for com- 
plete oxidation gives the value in terms of grams per cc, e.g., 1.4 gram of ore 
containing 69.2% Fe required a titration of 95 cc. KjCrjOr solution, then, 

(69.2X1.4) ^, ^^,^« 
1 cc. =^^ — — - — -'-^95 =0.0102 gram Fe. 
100 

stannous Chloride. Sixty grams of the crystallized salt dissolved in 600 
cc. of strong HCl and made up to one liter. The solution should be kept well 
stoppered. 

Mercuric Chloride. Saturated solution of HgCU (60 to 100 grams per liter). 

Potassium Ferricyanide, KtFe{CN)i. The salt should be free of ferrocyanide, 
as this produces a blue color with ferric salts, which would destroy the end- 
point. It is advisable to wash off the salt before using. A cr>'stal the size of 
a pinhead dissolved in 50 cc. of water is sufficient for a series of determinations. 
The solution should be made up fresh for each set of determinations. 

Apparatus. Chamber burette. Tliis should read from 75 to 90 cc. in 
tenths and from 90 to 100 cc. in twentieths of a cc. 

Test Plate. The usual porcelain test-plate with depressions may be replaced 
by a very simple and efficient test-sheet made by dipping a white sheet of paper 
in paraffin. The indicator does not cling to this surface, the drops assuming 
a spherical form, which renders the detection of the end-point more delicate. 

Procedure. Iron Ores. The amount of sample taken should be such that 
the actual iron present would weigh between 0.9 to 1.1 gram. This weight 
can be estimated by dividing 95 by the approximate percentage of iron present, 
e.g., for 50% Fe ore take f^ = 1.9 gram; 95% iron material would require 1 
gram, whereas 20% Fe ore would require 4.75 grams. 

For samples containing less than 20% Fe it is advisable to use N/10 K2Crj07 
solution. 

The sample should be finely ground (SO-mesh). 

Solution. The hydrochloric acid method for solution of the oxidized ore 
with subsequent carbonate fusion of the residue is recommended as being suitable 
for iron ores, briquettes, and materials high in iron. 

Reduction. HS reduction is recommended in ores containing arsenic or 
titanium. SnCl, in very slight excess, followed by mercuric chloride, HgCU, 
gives excellent results in absence of other reducible salts of elements, Cu, As, etc. 

Test Lead. The easy manipulation and efficiency of this method of reduc- 
tion makes it applicable for a large variety of conditions. The acid solution 
preferably, HCl, is diluted to about 150 to 200 cc, containing 15 to 20 cc con- 
centrated hydrochloric acid (sp.gr. 1.19). Sufficient test lead is added to cover 
the bottom of a No. 4 beaker. The solution covered is. boiled vigorously until 
it becomes colorless. Copper, if present, is precipitated, as well as platinun\, 



218 IRON 

and small amounts of arsenic and antimony eliminated from the solution during 

the reduction of the iron. The cooled solution is poured into a 600-cc. beaker 

and the mat of lead remaining in the No. 4 beaker washed free of iron, two or 

three 50-cc. washings being sufficient. The main solution and washings are 

combined for titration. If the solution is slightly colored, due to reoxidation 

of iron, a few drops of stannous chloride solution are added to reduce it, followed 

by an excess of HgClj solution, 20 to 25 cc, and allowed to stand five minutes. 

Titration, The standard potassium dichromate is run into the solution to 

within 5 to 10 cc. of the end-point, this having been ascertained on a pK)rtion 

of the sample. The dichromate is run in slowly near the end-reaction, and 

finally drop by drop until a drop of the solution mixed with a drop of potassium 

ferricyanide solution produces no blue color during thirty seconds. A paraffined 

surface is excellent for this test. 

FeXlOO 

Cc. KjCr207 multiplied by value per cc. =Fe present in sample. % = —: — . 

wt. taken 

Notes. If SnCls solution has been used for reduction of the iron, it is necessaiy 
to add the HgCU rapidly to a cold solution, as slow addition to a warm solution is 
apt to precipitate metallic mercury. 

In ca^e an excess of dichromate has been added in the titration, as often occurs, 
back titration may be made with ferrous ammonium sulphate (N 114)2804 -FeSOi-eHzO. 
N/IO solution of this reagent may be prepared by dissolving 9.81 grams of the clear 
crystals in about 100 cc. of water, adding 5 cc. of concentrated H2S04 and making to 
250 cc. The solution should be standai^ized against the dichromate solution to get 
the equivalent values, by running the dichromate directly into the ferrous solution. 

The ferricyanide indicator should be made up fresh each time it is required. 

Large amounts of manganese in the iron solution titrated cause a brown coloration, 
which masks the end-point. Nickel and cobalt, present in large amounts, are objection- 
ab'e for the same reason. This interference may be overcon e by using very dilute 
acid solutions of ferricyanide indicator, so that the insoluble ferricyanide of these 
metals will not form. 

Potassium Permanganate Method for Determination of Iron 

Introductioii. The method depends upon the quantitative oxidation of 
ferrous salts to the ferric condition when pK)tassium permanganate is added to 
their cold solution, the following reaction taking place: 

10FeSO4+2KMnO4+8H2SO4=5Fe2(SO4)3+K2SO4+2MnSO4-h8H2O. 

Hydrochloric acid in presence of iron salts has a secondary reaction upon 
the permanganate, e.g., 

2KMnO4+16HCl=2KCl-h2MnCl24-8H,O4-10Cl. 

This reaction may be prevented by addition of large amounts of zinc or man- 
ganous sulphates together with an excess of phosphoric acid.^ It is preferable, 
however, to expel HCl, when this has been used as a solvent, by adding sul- 
phuric acid and taking to fumes. The solution is diluted and reduced with zinc 
and titrated as directed. 

The reduction of ferric sulphate is best accomplished by passing the solution 
through a column of amalgamated zinc in the Jones redactor. In presence 
of titanium, reduction is accomplished by H2S in a hydrochloric acid solution of 

the iron. . 

* Jour. Am. Chem. Soc, 17, 405. 



IRON 219 

Since potassium permanganate enters into reaction with acid solutions of 
antimony, tin, platinum, copper and mercury, when present in their lower state 
of oxidation, (also with manganese in neutral solutions) and with SOj, H^S, N2O, 
ferrocyanides and with most soluble organic bodies, these must be absent from 
the iron solution titrated. 

Potassium permanganate produces an intense pink color in solution, so that 
it acts as its own indicator. 

Solutions Required. Standard Permanganate Solutions. As in case of 
pK)tassium dichromate, it is convenient to have two standard solutions, N/5 
and N/10. 

From the reaction given above it is evident that 2 KMn04 are equivalent to 
5 oxj'gens, e.g., 2KMn04=K20+2MnO-f 50, hence a normal solution would 
contain one-fifth of the molecular weight of KMn04 =31.6 grams of the pure salt. 
Hence a N/5 solution would contain 6.32 grams per liter and a N/10 solution 
3.16 grams. 

Since commercial potassium permanganate is seldom pure, it is necessary to 
determine its exact value by standardization. This is commonly accompUshed 
by any of the following methods: 

(a) By a standard electrolytic iron solution. 

(6) By ferrous salt solution, e.g., (NH4)2S04-FeS04-6H20. 

(c) By oxalic acid or an oxalate. 

Reaction. 2KMn04-h5Na2C204+8H2S04 

=K,SO4+2MnSO4+5Na^O4+10CO2+8H2O. 

Standardization of KMn04 against sodium oxalate is recommended as the 
most accurate procedure. The salt has no water of crystallization and is not 
hygroscopic. It can be obtained from the Bureau of Standards with a guarantee 
of purity. Traces of moisture can be expelled by heating the salt to 120° C. for 
two hours, then cooling in a desiccator. 

N/5 Na2C204 contains 13.40 grams per liter, N/10 solution contains 6.7 
grams. For standardization of N/5 KMn04, 3.35 grams of the sodium oxalate 
are dissolved in warm (70° C.) water (about 200 cc), 50 cc. of 2N. H2SO4 are added 
and the solution made up to 250 cc. 

N/5 KMn04 contains 6.32 grams of the salt per liter. It is advisable to 
dissolve 6.4 grams of the salt in about 500 cc. of hot water and filter the solution 
through asbestos to remove any dioxide of manganese that may be present, 
as MnOj aids in the decomposition of KMn04 solution. The reagent should be 
kept tightly sealed in a dark bottle well protected from the light. This solution 
should stand two or three days before standardization. 

To standardize the solution 100 cc. of the N/5 sodium oxalate solution is heated 
to about 70° CJ. and the permanganate solution added from a 100-cc. burette 
very slowly in small portions at a time, allowing the color to fade after each 
addition before adding more. When within 6-10 cc. of the end-point the per- 
manganate solution should be added drop by drop until a faint permanent pink 
color persists. 

Procedure for the Determination of Iron by the Jones Reductor 

Preparation of Sample. Such an amount of the sample is taken that the 
iron content is between two- and three-tenths of a gram (0.2 to 0.3 gram). If 
hydrochloric acid has been required to effect solution, or hydrochloric acid and 



220 



IRON 



nitric acid (25 cc. : I cc), as in case of iron and steel, 4 to 5 ec. cone, sulphuric 
ar3 added, and the solution evaporated to small bulk on the steam bath and to 
SOi fumes to remove hydrochloric acid. The iron is taken up with about 50 co. 
dilute sulphuric acid, 1 : 4, heating if necessary, and filtering if an insoluble 
residue remains. 

Preparatioii of the Reductor. Cleaning out the apparatus. See Fig. 40. 
The stop-cock of the reductor is closed, a heavy-walled flask or bottle is put into 

position at the bottom, and 50 cc. of dilut« 

sulphuric acid poured into the funnel. The 
cock is opened and the acid allowed to flow 
slowly through the zinc m the tube, applying 
a gentle suction. Before the acid has drained 
out of the funnel, 50 ce. of water are added, 
followed by 50 cc. more of dilute sulphuric 
acid and 50 cc. of water in turn. The stop- 
cock is turned off before the water has drained 
completely from the funnel so that the zinc is 
always covered by a solution of acid or water. 
This precaution should be olraerved in all 
determinations with the Jones reductor to pre- 
vent the inflow of air into the column of zinc. 
The contents of the flask being emptied and 
the flask replaced, the apparatus is ready for 
the detomkination of the blank. 

Detenoinatioii of the Blank. Fifty cc. 
of dilute sulphuric acid, 1:4, arc passed 
through the reductor, followed by 250 cc. of 
' distilled water, according to the directions 
given above. The acid solution in the flask 
is then titrated with K/IO KMnO, solution. 
If n>ore than 3 or 4 drops of the permanga- 
nate are required, the operation must be re- 
peated until the blank titration does not 
exceed this amount. The final blank obtained 
should be deducted from the regular determinations for iron. Tlie end-point of 
the titration is a famt pink, persisting for one minute. 

Reduction and Titration of the Iron Solution. The sample is diluted 
to 200 cc, and, when cold, is run into the funnel, the atop-cock opened and 
the solution dra^Ti slowly through the colunm of zinc into the flask, about four 
minutes being re<!uired for 200 cc. of solution. Before the funnel has com- 
pletely drained, rinsings of the vessel which contained the sample are added; 
two 50-cc. portions are sufficient, followed by alKiut 50 cc of water. The 
stop-cock is closed Iwfore the solutions ha\c complctoh draniod from the funnel. 
Titration. The flsisk is reniovetl and tenth normal solution of perman- 
ganate added until a faint pink color, perM'.tnig one minute is olitained. The 
olank is deducted from the cc. reading of the burette 

Cc. KMnOi thus found multiplied b\ the ^nluc of the reagent in terms of 
N/10 =true value of N/IO KM11O4 required to oxidize the reduced iron. 

One cc. X/IO KMnO, = .005584 gram Fe; or .007984 gram Fe,0,. 







FiQ. 40 — Jones Reductor. 



[JIRON 221 

This weight, divided by the weight of the sample taken, and multiplied by 
100=per cent iron, or iron oxide in the sample, accordiQg to the factor taken 
above. 

Stannous Chloride Method for Determination of Ferric Iron 

The procedure is based upon the reduction of the yellow ferric chloride to 
the colorless ferrous salt by stannous chloride, the following reaction taking place : 

2FeCl,+SnCl, =2FeCI,+SnCl,. 

The method is of value in estimating the quantity of ferric iron in presence 
of ferrous, where the two forms are to be determined. In order to obtain the 
total iron the ferrous is oxidized by adding 
a tew crystals of potassium chlorate and tak- 
ing to dryness to expel chlorine, and then 
titrated with stannous chloride. 

The accuracy of the method depends 
upon the uniformity of conditions of tem- 
perature, concentration, etc., of making the 
run with the sample and of standardizing the 
stannous chloride. The solution should be 
free from other oxidizing agents, or from 
salts that give colored solutions. 

The amount of iron in terms of ferric 
oxide that can be estimated by this procedure 
ranges from 0.002 gram to 0.05 gram. 

Reagents. Stannoua Chloride Solution- 
The reagent is prepared by dissolving 2 grams 
of stannous chloride crystals in hot concen- 
trated hydrochloric acid and making up to 
1 liter. The solution should Iw kept in a 
dark bottle to which the titrating burette is 
attached in such a way that the liquid may 
be siphoned out into this, as shown in the 
illustration, Fig. 41. The air entering the 
bottle passes through phosphorous or pyro- 
gallic acid to remove the oxygen. In this way, protected from the air, the 
reagent will keep nearly constant for several weeks. It is advisable, however, to 
restandardize the solution alwut every ten to fifteen days. One cc. will be equiv- 
alent to about 0.001 gram Fe. 

Standa-d Iron Solution. 8.C322 grams of ferric ammonia alum is dis.solved 
in dilute hydrochloric acid and made up to one Lter. The iron is determined in 
100-cc. portions by the dichromate meUiod. One cc. will contain about O.COl 
gram Fe. 

Procedure. To the sample in a casserole is added 25 cc. of concentrated 
hydrochloric acid and an equal volume of water. The resulting solution is 
heat«d to boiling and quickly titrated with the stannous chloride reagent, 
until the yeUow color fades out and the solution becomes colorless. 

Note. The titration should be done quickly, as t 
and tie solution again become yellow. The true e 
oolorieiB solution. 




Fig. 41. — Apparatus for Stannoua 
Chloride Titration of Iron. 



222 IRON 

COLORIMETRIC METHODS FOR THE DETERMINATION OF 

SMALL AMOUNTS OF IRON 

Iron Traces. Sulphocyanate (Thiocyanate) Method^ 

Introduction. By this method 1 part of iron may be detected in 50 million 
parts of water. The presence of free mineral acid increases the sensitiveness 
of the method, so that it is especially applicable to the determination of small 
amounts of iron in mineral acids. It is available in presence of many of the 
ordinary metals and in presence of organic matter. Silver, copper, cobalt, mer- 
curic chloride, however, interfere. 

Nitric acid gives a color with sulphocyanates that may be mistaken for 
iron. 

This method, like the stannous chloride method, determines only the ferric 
iron. It is based on the fact that ferric iron and an alkali sulphocyanate, 
ammonium or potassium sulphocyanates, in an acid solution gives a red color, 
the intensity of which is proportional to the quantity of iron present. The 
color is due to the formation of the compound, Fe(CNS)3-9KCNS -41120. 

Reagents Required. Standard Iron Solution. A ferric solution, the iron 
content of which has been determined, is diluted and divided so as to obtain 
0.0004 gram Fe. This is made up to 2 liters with water containing 200 cc. of 
iron-free, C.P. H2SO4. One hundred cc. of this solution, together with 10 cc. 
of normal ammonium sulphocyanate solution, is used as a standard. One 
himdred cc. contains 0.00002 gram Fe. 

Normal sulphocyanate contains 76.1 grams of NH4CNS per liter. 

Procedure. The weighed sample, 1 to 10 grams, or more if necessary, is 
dissolved in dilute H2SO4 and oxidized by adding dilute permanganate, KMn04, 
solution drop by drop until a faint pink color is obtained. The sample is diluted 
to exactly 100 cc. and is poured into a burette graduated to -^ cc. Two 
colorless glass cylinders of the lOO-cc, Nessler type are used for comparison 
of standard and sample. Into one cyUnder is poured 100 cc. of the standard 
solution, made as directed above. Into the second cylinder containing 10 cc. 
of sulphuric acid with 10 cc. ammonium sulphocyanate, NH4CNS, diluted to 
60 or 70 cc, the sample is run from the burette until the depth of the color 
thus produced on dilution to 100 cc. exactly matches the standard. From the 
number of cc. used the weight of the sample is calculated. One hundred cc. 
of the standard contains 0.00002 gram Fe. 

Dividing the weight of iron in the standard by the weight of sample used 
and multiplying by 100 gives the per cent of iron in the sample. 

Notes. If other metals are present, that form two scries of salts, they miist be in 
the higher sta^e of oxidation, or the color is destroyed. (Sutton.) Oxalic acid, if pres- 
ent, destroys the color. Oxidation with KMn04 or KCIOa with subsequent removal of 
CI2 prevents this interference. (Lunge, C. N., 73, 250.) 

Chlorides of the alkaline earths retard or prevent the sulphocyanate reaction. 
(Weber, C. N., 47, 165.) 

The colorimeter used for the determination of minute quantities of lead would serve 
admirably for the determination of traces of iron by the sulphocyanate method. 

Acids, hydrochloric or sulphuric (diluted), may be added directly to the ammonium 
sulphocyanate solution. 

1 Thomson, J. C. S., 493, 1885, and C. N., 61, 259. Kruss and Moraht, C. N., 64, 
255. Dayies, C. N., 8, 163. 



IRON 



223 



Salicylic Acid Method for Determining Small Amounts of Iron ^ 

Salicylic acid produces an amethyst color with neutral solutions of ferric 
salts, the depth of the color being proportional to the concentration of the ferric 
iron in the solution. The reaction is useful in determining small amounts of 
iron in neutral salts, such as sodium, ammonium, or pK)tassium alums, sulphates, 
or chlorides, zinc chloride, etc. Phosphates, fluorides, thiosulphates, sulphites, 
bisulphites and free mineral acids should be absent. The sample should not 
contain over 0.0002 gram iron, as the depth of color will then be too deep for 
colorimetric comparisons. As low as 0.00001 gram ferric iron may be detected. 
Ferrous iron produces no color with the reagents, hence the procedure serves 
for determining ferric iron in presence of ferrous. 

The material is dissolved in 20 cc. of pure water, the sample filtered if cloudy, 
and transferred to a Nessler tube. Dilute potassium permanganate solution 
is added until a faint pink color is produced and then 5 cc. of a saturated solution 
of saUcylic acid. (The reagent is filtered and the clear solution used.) Com- 
parison is made with standard solutions containing known amounts of ferric 
iron, the standards containing the same reagents as the sample. If desired 
the standard iron solution (0.086 gram ferric ammonium alum, clear crystals, dis- 
solved in water containing 2 cc. of dilute sulphuric acid and made to 1000 cc, 
each cc. contains approximately 0.00001 gram Fe'") is added from a burette 
to 5 cc. of salicylic acid diluted to 25 cc. in a Nessler tube, until the color of the 
standard matches the sample. A plunger is used to stir the liquids. 




TECHNICAL ANALYSIS OF IRON AND STEEL 

The elements carbon, manganese, phosphorus, sulphur, and silicon are in- 
variable constituents of iron and steel, and are always included in an analysis. 
Copper and arsenic are sometimes found; aluminum, chro* 
mium, nickel, molybdenum, tin, titanium, tungsten, vana- 
dium, and zinc occur in special alloy steels. Minute traces of 
oxygen, hydrogen, and of many other elementary constituents 
frequently are present, but are of so little importance that 
they are seldom considered in an analysis. 

Our attention is drawn in this chapter to the more impor- 
tant constituents, whose estimation is required in the daily 
routine analysis of a steel works laboratory. The elements 
considered are carbon — carbide or combined carbon and graph- 
itic carbon, manganese, phosphorus, sulphur, and silicon. 
Determination of the elements of special alloy steels contain- 
ing aluminum, chromium, nickel, titanium, tungsten, vana- 
dium, etc., are given in the chapters on the elements in question; 
for example the determination of vanadium in steel will be 
found in the chapter on Vanadium, chromium and copper in 
the chapters on Chromium and Copper, etc. 

As is generally the case, a large number of determinations 
are required in the steel works laboratories and it is not an Diyj^jing Pipette 
uncommon thing for one man to turn out 60 to 100 determina- 
• tions a day. To accomplish this, simple and rapid procedures are required. 
When the metal is unusually high in an undesirable constituent it is indicated 
* Method of W. S. Allen, by courtesy of the General Chemical Company. 



15 
cc 




Fig. 42. 



224 IRON 

by the test, and a confirmation of the result is obtained by an additional test, 
exercising extreme care, and using a procedure giving results of the highest 
accuracy. Fortunately the analysis of steel has received considerable attention 
and rapid methods have been worked out which are extremely accurate. 

The procedures briefly outlined have proven of value to analysts of iron 
and steel. While in charge of the laboratory at Baldwin Locomotive Works, 
the author found that a skilled analyst was able to turn out 125 determinations 
of combined carbon, or 100 of manganese, or of sulphur, or 50 determinations 
of phosphorus, or 25 determinations of silicon per day by the procedures given. 
This necessitates the use of a large number of beakers and flasks, ample desk 
room, individual balances, hot plates, and hoods to acconunodate a dozen to 
two dozen beakers or flasks at a time, and a carefully planned system. 

The dividing pipette, shown in Fig. 42, is useful for adding a definite amount 
of reagent to the sample. 

In addition to the short methods, we include the procedures recommended 
by the U. S. Bureau of Standards, for cases where accuracy is essential and 
time a secondary consideration. 

Preparation of the Sample 

The metal is sampled by drilling with a clean twist drill, using no water 
or oil. 

Hard grades of pig iron, chilled iron, ferromanganese, quenched steel, etc., 
are broken down to a coarse powder in a chilled steel mortar. 

Combined or Carbide Carbon— Colorimetric Method 

Rapid Method. 0.2 gram of well-mixed drillings is placed in a test-tube 
6Xi ins. and 4 to 10 cc. of nitric acid (sp.gr. 1.2) added from a burette, the 
test-tube being placed in cold water to prevent too violent action. The amount 
of acid added is governed by the carbon-content of the steel (see chapter on 
Carbon, page 108). After the violent action has ceased, the tube is placed in a 
specially-designed water bath, the water heated to boiling and boiled for twenty 
minutes or more until the solution in the tube has become perfectly clear. 
The sample is now removed, washed into a color carbon tube and compared with 
a standard steel of the same class of material as that examined. Full details of 
the procedure may be found in the chapter on Carbon. 

Iron and steel containing graphite must be filtered before making comparisons. 
The solution, diluted with one-half its volume of water, is filtered through a small 
filter paper into a test-tube. The residue is washed with a fine jet of distilled 
water until free of color. The filtrate is compared with a standard sample of 
similar composition treated in the same way. 

Steel containing chromium, copper, nickel, and elements yielding a colored 
solution should not be examined by the colorimetric methods. 

Method of the Bureau of Standards. Total and graphitic carbon are 
determined and the difference taken as combined carbon. 



IRON 225 

Specifications for Combined Carbon. 

MateriaL Per eent Carbon. 

Boiler rivets . 15 

Seamless boiler tubes 0. 18-0. 25 

Boiler and fire-box plates 0. 15-0.25 

Cylinder grade pig iron 0.25-0.50 

Forged and rolled steel wheels 0. 70-0. 75 

Steel blooms for forgings, open hearth, basic 0.40 0. 55 

Steel blooms for forgings, acid . 35-0 . 40 

Bolt steel 0.22-0.28 

»Spring steel 0.90-1 . 10 

Cfrank axles forged 0.35-0.55 

Casting not over . 35 

Tire steel 0.65-0.75 

Floor grade pig iron not over . 40 

Total Carbon 

The determination is required for an accurate estimation of carbon where 
the color test indicates the carbon content outside the limits of requirement, 
or in cases where interfering substances are present. In material where the 
carbon content is of extreme importance, the color method is not used. Details 
of the procedure for determining carbon by direct combustion are given in 
the chapter on Carbon. The following procedure is recommended by the 
Bureau of Standards: 

(a) In Irons. Two grams of iron are mixed with about twice the weight 
of purified ferric oxide. The mixture is placed in a platinum boat, which is 
lined with a suitable bed material, and is burned in a current of oxygen, as 
described below. 

(6) In Steels. The method is the same as for irons with omission of the 
ferric oxide mixture. 

Details of Direct Combustion Method. Furnaces and Temperature of 
Burning. Porcelain tubes wound with ''nichrome" wire, provided with suit- 
able heat insulation and electrically heated, are used, and readily give tem- 
peratures to 1100° C. Type FB 301 Hoskins tube furnace and the hinged 
type, Fig. 326, are satisfactory. The temperature control is by means of an 
ammeter and rheostat in series with the furnace, with occasional check by a 
thermocouple. 

Boats and Lining, Platinum boats provided with a long platinum wire for 
manipulation in the tube are mostly used ; alundum ones occasionally. The bed 
or lining on which the steel rests is 90-mesh *^RR alundum, alkali-free, specially 
prepared for carbon determination." A layer of this alundum is also placed 
in the bottom of the combustion tube to prevent the boat sticking to the glaze. 
A platinum cover for the boat is sometimes used, and is essential when the 
combustion is forced. 

The nature and quality of the bed material are matters of great importance. 
Alumina as prepared from the sulphate or from alum may not be free from 
sulphate or alkali, both of which have given serious trouble at the Bureau. The 
alkali, if present, may not manifest itself by an alkaline reaction until after 
one or two combustions have been made, using the same bed material. Even 
the ordinary white "alundum" on the market carries a few hundredths of 1% 
of alkali. Iron oxide has been tried, and when pure should, apparently, give good 



226 lEON 

service. As yet, however, it has been difficult to obtain or prepare acceptable 
material for use with steels. Quartz sand gives rise to a fujible slag, which 
melting before combustion is complete, incloses bubbles of carbon dioxide gas. 
This defect would probably inhere in any other material of an acid character. 
The presence in the silica bed after combustion of crystals which appear to be 
carborundum, have occasionally been noted.^ 

Purity of Oxygen. Blanks. The Bureau makes its oxygen electrolytically, 
and its content of this element is usually 99 to 99.5%, and sometimes higher. 
Even with this gas a slight blank is usually obtained. When running a blank, 
in additional to the usual precautions, the rate at which the oxygen is intrcK 
duced should be the same as when burning a sample, and the time should be 
three to five times as long. 

Method of Admitting Oxygen and Rate of Combustion. The furnace being 
at the proper temperature, the boat containing the sample is introduced. Oxygen 
is admitted either at once or after the boat has reached the temperatiu^ of the 
furnace, as the operator prefers, or as the nature of the steel may demand. 
The rate of flow of the oxygen varies with the absorption apparatus used and 
with the preference of the operator, and may be considerably more rapid "~^hen 
absorbing carbon dioxide in soda lime than in an alkaline solution. A apid 
flow of oxygen also facilitates the burning of resistant samples. A continuous 
forward movement of the gas current is maintained at all times. The time 
for a determination varies, of necessity, with the nature of the sample and the 
rate of flow of the oxygen, ranging from ten to thirty minutes. The endeavor 
is to obtain a well-fused oxide. With all samples close packing in a small space 
is conducive to rapid combustion and to fusion of the resulting oxide. 

Authorities differ as to the advisabiHty of allowing the oxide of iron to fuse 
thoroughly. Even when fusion does take place additional carbon dioxide is 
obtained very frequently by grinding the oxide and rebuming. Often more 
than one regrinding and rebuming is necessary in order to reduce the amount 
of carbon dioxide obtained to that of the constant blank. 

Oxides of sulphur have been found very difficult to eliminate from the gases 
leaving the tube. Lead peroxide ("nach Dennstedt") heated to 300° C. and 
zinc at room temperature appear to retain them best 

Attention is called to the inadmissibility of using dry agents of different 
absorptive power in the same train, in positions where a difference could possibly 
affect results. 

Weighing of Tubes. There is much greater difficulty in securing constant 
conditions when weighing absorption tubes than is usually considered to be 
the case. Electrical effects, caused by wiping as a preliminary to weighing, 
may occasionally cause errors in weight running into the ndUigrams. The 
use of counterpoises of equal volume and similar material and shape is recom- 
mended. 

If tubes are weighed full of oxygen, care is necessary to secure a uniform 
atmosphere in them. Even though the attempt is made to keep the apparatus 
alwa>'s lull of oxygen, some air is admitted when the boat is pushed into the 
combustion tube, and a much longer time is required to displace this than is 
usually allowed, imless the flow of oxygen during aspiration is rapid. The 
same is true if the tubes are weighed full of air by displacing the oxygen left in 
them after the steel is burned. Another source of error may arise from the 

^ Statement of Mr. George M. Berry, of the Halcomb Steel Co.] 



IRON 227 

air admitted when putting the boat into the tube, if this air contains much 

carbon dioxide, as is the case when a gas furnace is used. The boat is usually 
pushed at once into the hot furnace, and as combustion begins almost imme- 
diately, there is no opportimity for displacing this air before the steel begins 
to bum. 

Graphite in Iron 

Two grams of iron are dissolved in nitric acid (sp.gr. 1.20), using 35 cc. and 
heating very gently. The residue is collected on an asbestos felt, washed with 
hot water, then with a hot solution of potassium hydroxide (sp.gr. 1.10), fol- 
lowed by dilute hydrochloric acid and finally by hot water. After drying at 
100® C, the graphite is burned in the same manner as the total carbon, but 
without admixture of ferric oxide. 

Manganese in Iron and Steel. Ammonium Persulphate Method 

Small amounts of manganese may be determined colorimetrically by the 
persulphate method, provided the sample does not contain over 1.5% of man- 
ganese. The procedure given in detail in the chapter on Manganese, page 267, 
in brief is as follows: 

Reaction. 2Mn(NO,),+5(NH4)tS,08-h8H,0 

-5(NH4),S04+5H^04+4HNO,+2HMn04. 

0.1 to 0.2 gram of steel, according to the amount of manganese in the sample, 
is placed in a 10-in. test-tube and 10 cc. of nitric acid (sp.gr. 1.2) are added. 
The sample is heated in a water bath until the nitrous fumes are driven off and 
the steel is completely in solution, fifteen cc. of silver nitrate solution are 
added to the cooled sample, followed immediately with about 1 gram of ammo- 
nium persulphate crystals. The solution is warmed (80 to 90° C.) imtil the 
color commences to develop, and then for half a minute longer, and then placed 
in a beaker of cold water until the solution is cold. Comparison is now made 
with a standard steel treated in the same way. The comparison being made 
exactly as indicated for determining carbon by the color method. See chapter 
on Carbon. 

Example. If the standard, containing 0.6% Mn is diluted to 15 cc, each 
cc. =0.04% Mn. If the sample required a dilution of 20 cc. to match the 
standard, then 0.04X20 =0.8% Mn. 

Lead Oxide Method ( Deshey) • 

Oxidation of the manganese in the steel is effected in a nitric acid solution 
by addition of red lead (or by lead peroxide) ; the lead peroxide, formed oxidizes 
the manganese nitrate to permanganic acid. The solution is now titrated with 
standard sodium arsenite, the following reaction taking place: 

2HMn04+5Nai»AsO,+4HNO,=5Na,As04+3H,0+2Mn(NO,),. 

0.5 gram of steel is placed in a 150-cc. beaker and dissolved with about 
80 cc. of nitric acid (sp.gr. 1.12). After violent action has subsided, the beaker 



228 IRON 

is placed on a hot plate and when the iron has dissolved, 20 cc. of water added. 
The manganese is now oxidized by adding red lead in small portions at a time, 
until the solution appears brown with a pinkish purple foam on the surface. 
The solution is diluted with hot water until the volume is about 100 cc. and 
then boiled for a few minutes. It is now placed in a dark closet to cool. (A 
fresh batch of samples may be started in the meantime.) The solution is 
carefully decanted off from the peroxide, and with the washings of the peroxide 
residue, titrated with standard sodium arsenite to the yellowish green end- 
point. The sodium arsenite is made by dissolving 4.96 grams of pure arsenous 
acid together with 25 grams of sodium carbonate in 200 cc. of hot water and the 
solution diluted to 2500 cc. The arsenite is standardized against a steel sample 
of known manganese content, or against standard permanganate solution. 

Bismuthate Method for Determining Manganese, Recommended 

by the U. S. Bureau of Standards^ 

This is the most accurate method for determining manganese in iron and 
steel. The procedure is as follows: 

Procedure. One gram of drilUngs is dissolved in 50 cc. of nitric acid 
(sp.gr. 1.135) in a 200-cc. Erlenmeyer flask. Irons should be filtered. The 
solution is cooled, about 0.5 gram of sodium bismuthate is added, and it is 
then heated until the pink color has disappeared. Any manganese dioxide 
separating is dissolved in a slight excess of a solution of ferrous sulphate or 
sodium sulphite. The solution Is boiled till free from nitrous fumes. After 
cooling to 15° C, a sUght excess of bismuthate is added and the flask is shaken 
vigorously for a few minutes. Then 50 cc. of 3% nitric acid is added and the 
solution is filtered through asbestos. A measured excess of ferrous sulphate 
is run in and the excess titrated against permanganate solution which has been 
compared with the iron solution on the same day. A great many steels now 
carry small amounts of clu*omium as impurity. In such cases titration against 
arsenite solution is recommended, or removal of the chromium by zinc oxide and 
subsequent determination of the manganese by the bismuthate method. 

Permanganate solutions are standardized against sodium oxalate (Bur. Stds. 
Sample No. 40) as prescribed by McBride.* 

Specifications for Manganese in Iron and Steel 

Material. Percentage of Manganese. 

BoUcr rivets 0.30-0.60 

Boiler and fire-box plates not over 0. 45 

Seamless boiler tubes . 40-0 . 65 

Floor-grade pig iron not over . 80 

Cylinder-grade pig iron 0.50-0.80 

Forged and rolled steel wheels 0.60-0.80 

Steel blooms lor forgings not over . 70 

Bolt steel not under . 50 

Spring steel not over . 50 (0 . 25 desired) 

CStmk axles, forged not over . 75 

Castings not over . 75 

Tire steel 0.50-0.75 

iSee page 263. 

« Bull. Bur. Sds., 8, 641. J. Am. Chem. Soc, 84, 393, 1912. 



IRON 229 

Determination of Phosphorus 

The procedures outlined by the Bureau of Standards are generally used in steel 
works laboratories. 

(a) Preparatioii of Solution and Precipitation of Phosphorus. Two 
grains of sample are dissolved in nitric acid (sp.gr. 1.135) and the solution is 
boiled until brown fumes no longer come off. Ten cc. of permanganate solu- 
tion (15 grams to 1 liter) are added, and the boiling is continued. Sodium sul- 
phite solution is added to dissolve the oxide of manganese, and the solution is 
again boiled and then filtered. With irons the insoluble residue should be tested 
for phosphorus. After cooling the filtrate, 40 cc. of anmionia (sp.gr. 0.96) are 
added, the solution is agitated, and when the temperature is at 40° C, 40 cc. 
of molybdate solution ^ are added and the solution is shaken vigorously for 
five minutes. After setthng out, the yellow precipitate is treated according to 
one of the following methods, 6 or c: 

(6) Alkalimetric Method. The precipitate is washed with 1% nitric acid 
solution followed by 0.1% potassium nitrate solution until the washings are 
no longer acid. The precipitate is dissolved in a measured excess of standard- 
ized sodium hydroxide solution and titrated back with standardized nitric acid 
using phenolphthalein. The solutions are standardized against a steel with 
a known amount of phosphorus. 

(c) Molybdate Reduction Method. The precipitate is washed ten to 
fifteen times with acid ammonium sulphate (prepared according to Blair) or 
until the washings no longer react for iron or molybdenum. It is dissolved 
in 25 cc. of ammonia (5 cc. ammonia of 0.90 sp.gr. to 20 cc. water). The filter 
is washed well with water and 10 cc. of strong sulphuric acid added to the 
filtrate, which is run through the reductor at once and titrated against a N/30 
permanganate solution which has been standardized against sodiiun oxalate, 
as prescribed by McBride.* 

Specifications for the Amount of Phosphorus 

Class of Material. v. ^^"iW ^^ ^^°^' 

phorus Allowed, per cent. 

Boiler steel 0.05 

Forged and rolled steel wheels . 05 

Steel blooms for forgings . 05 

Crank axles ,. . 05 

Tire and bolt steel 0.05 

Spring steel . 05 

Spring steel desired . 03 

Fire-box plates . 03 

Castings . 06 

Floor-grade pig iron . 5-0 . 9 

Cylinder iron 0.5-0.9 

Determination of Sulphur 

Rapid-evolution Method. Volumetric. Five grams of iron or steel are 
placed in a 500-cc. Erlenmeyer flask, provided with a two-holed rubber stopper, 
through which passes a long-stem thistle tube reaching to the bottom of the 

* Blair, "Chemical Analysis of Iron," (7th Ed.), p. 97. 

« Bull. Bur. Stds., 8, 641. J. Am. Chem. Soc, 84, 393, 1912. 



230 IRON 

flask, and a delivery-bulb condenser, connected by means of a rubber tube to 
an alDSorption bulb. (See sketch of apparatus in the chapter on Sulphur, volu- 
metric methods, page 399.) 

About 25 to 35 cc. of an ammoniacal solution of cadmium chloride are placed 
in the absorption bulb, the apparatus connected and about 100 cc. of dilute 
hydrochloric acid (sp.gr. 1.1) poured through the thistle tube into the flask 
•containing the drillings. The mixture is heated gently until the sample goes 
into solution and then boiled until steam escapes from the apparatus. The 
reagent in the absorption bulb should remain alkaline, otherwise a loss of sulphur 
is apt to occur. 

The absorption bulb is now disconnected and the contents emptied into a 
400-cc. beaker and the bulb washed out with dilute hydrochloric acid after 
first rinsing out once or twice with water. The solution is now diluted to 
about 300 cc, and if not already acid, is made so by addition of more hydrochloric 
acid. 

Two to 3 cc. of starch indicator are added and the mixture titrated 
with standard iodine, stirring constantly during the titration. A permanent 
blue color in the end-pK)int sought. If much cadmium sulphide is present 
additional hydrochloric acid may be required. 

The number of cc. of iodine solution required multiplied by the factor of 
iodine to sulphur gives the amount of sulphur present in the sample taken. 

Notes. For a more complete description of the procedure see chapter on Sulphur. 

With certain pig irons low results are apt to be obtained by the evolution method. 
For such the gravimetric method given is recommended. 

Gray iron will evolve all its sulphur as HjS, white iron, gray water-chilled iron, 
gives up only part of its sulphur by the evolution method. The method gives low 
results for high carbon steel. 

In place of absorbing the HjS in cadmium chloride, the Bureau of Standards 
recommends absorption in an ammonicaal solution of hydrogen peroxide (5 cc. HjOj 
3% +25 cc. NHiOHj sp.gr., 0.90). The sulphuric acid formed is precipitated from 
a slightly hydrochloric acid solution, by barium chloride and weighed as BaSOi. 



Method by the U. S. Bureau of Standards. Gravimetric 

Sulphur by Oxidation 

Five grams of iron or steel are dissolved in a 400-cc. Erlenmeyer flask, using 
50 cc. of strong nitric acid. A little sodium carbonate is added, the solution 
is evaporated to dryne.<v?, and the residue baked for an hour on the hot plate. 
To the flask 30 cc. of strong hydrochloric acid are added, and the evaporation 
and baking are repeated. After solution of the iron in another 30 cc. of strong 
hydrochloric acid and evaporation to a sirupy consistency, 2 to 4 cc. of the 
same acid are added, followed by 30 to 40 cc. of hot water. The solution is then 
filtered and the residue washed with hot water. The sulphur is precipitated 
in the cold filtrate (about 100 cc.) with 10 cc. of a 10% solution of barium 
chloride. After forty-eight hours the precipitate is collected on a paper filter, 
washed first with hot water (containing 10 cc. of concentrated hydrochloric acid 
and 1 gram of barium chloride to the liter) until free from iron and then with 
hot water till free from chloride; or, first with cold water, then ^nth 25 cc. 
of water containing 2 cc. of concentrated hydrochloric acid to the liter. The 



IRON 231 

washings are kept separate from the mam filtrate and are evaporated to recover 
dissolved bariimx sulphate. 

With irons the paper containing the insoluble residue above mentioned is 
put into a platinum crucible, covered with -sodium carbonate free from sulphur, 
and charred without allowing the carbonate to melt. The crucible should 
be covered during this operation. Sodium nitrate is then mixed in and the 
mass fused with the cover off. An alcohol flame is used throughout. The melt 
is dissolved in water and evaporated with hydrochloric acid in excess to dryness 
in porcelain. The evaporation with water and hydrochloric acid is repeated 
to insure removal of nitrates. The '•esidue is extracted with a few drops of 
hydrochloric acid and water, the insoluble matter is filtered off, and bariiun 
chloride is added to the filtrate. The barium sulphate obtained is added to 
the main portion. 

Careful blanks are run with all reagents. 

Specifications for Sulphur in Iron and Steel 

Material. Specificationa, 

per cent. 

Seamless boiler tubes must be below . 05 

Cylinder iron 0. 05 

Forged and rolled steel wheels 0. 05 

Steel blooms for forgings,- basic and acid open hearth . 05 

Bolt steel 0.05 

Spring steel . 05 

Cfrank axles . 05 

Tire steel 0.05 

Castings 0.06 

Boiler rivets 0.04 

Floor-grade iron 0. 04 

Boiler plates 0.035 

Fire-box plates 0.035 

Amount desired in spring steel 0.030 

Muck bar iron . 02 

Determination of Silicon 

One gram of pig iron, cast iron, and high silicon iron, or 5 grams of steel, 
wrought iron, and low silicon iron are taken for analysis. (By taking multiples 
of the factor weight 0.4693, SiOj to Si, the final calculation is simplified.) The 
sample is placed in a 250-cc. beaker and 20 to 50 cc. of dilute nitric acid added. 
If the action is violent, cooUng the beaker in water is advisable. When the 
reaction subsides, 20 cc. of dilute sulphuric acid, 1:1, are added, the mixture 
placed on the hot plate and evaporated to dense white fumes. The residue 
is taken up with 150 cc. of water containing 2 to 5 cc. of sulphuric acid and 
heated until the iron completely dissolves. 

The solution is filtered and the silica residue washed first with hot dilute 
hydrochloric acid, sp.gr. 1.1, and then with hot water added in small portions 
to remove the iron sulphate. The residue is now ignited and weighed as silica. 

Note. If the ash is colored by iron oxide, silica is determined by diflference, 
after expelling the silica by adding 4 to 5 cc. of nydrofluoric acid and a few drops of 
sulphuric, taking to dryness and igniting he esidue. 

The following acid mixtures are recommended )y the U. P. Ry. For steel, wrought 
iron and low sihcon iron, 8 parts by volume of HNO3, sp.gr. 1.42; 4 parts of cone. 
HfSO*, 8p.gr. 1.84; 6 parts HCl, sp.gr. 1.2 and 15 parts by volume of water. For 



232 IRON 

dissolving pig iron^ cast iron and high silicon iron, a mixture of 8 parts by volume 
of strong nitric acid and 5 parts of strong sulphuric acid, diluted witn 17 volumes of 
water is used. 

Rapid Method for Determining Silicon in Foundry Work. Liquid iron, 
dropped into cold water from a ladle 3 ft. above the water, will form shot shaped 
according to forms resulting from its chemical constitution, silicon being an 
important factor. Round shot, concave upper surface, i to f in. in diameter, 
indicate over 2% silicon. Flat, or irregular shot indicate low silicon. Shot with 
elongated tails indicate very low silicon. . 

Method of the U. S. Bureau of Standards for Silicon 

The insoluble residue obtained in preparing the iron or steel for the gravi- 
metric sulphur determination is filtered off, ignited in platinum, and weighed. 
Evaporation with a little hydrofluoric acid and 1 drop of sulphuric acid and 
subsequent ignition gives by the loss of weight silica corresponding to the 
silicon of the sample. 

Specifications for Silicon in Iron and Steel 

Material. Amount of Silicon, per cent. 

Boiler and fire-box plates not over . 03 

Floor-grade pig iron 2 . 25-2 . 75 

Cylinder-grade pig iron 1 . 25-1 . 60 

Forged and rolled steel wheels not over . 20 

Spring steel not over . 25 (0.15 desired) 

Tire steel under 0.25 



LEAD 

Wilfred W. Scott 

Pb, at.wt. 207.2; sp.gr. 11.34; m.p. 327**; b.p. 1525* C; oxides, PbO, 

PbOj, Pb304. 

DETECTION 

Hydrochloric acid precipitates lead incompletely from its cold solution as 
white PbCU, soluble in hot water by which means it is separated from mercurous 
chloride and silver chloride. PbCh forms needle-like crystals upon cooling the 
extract. 

Hydrogen sulphide precipitates black PbS from slightly acid solutions along 
with the other elements of the group.* Yellow anunonium sulphide, sodium 
sulphide and the fixed alkalies dissolve out arsenic, antimony and tin. The 
sulphide of lead, together with bismuth, copper and cadmium, dissolve in hot 
dilute nitric acid, leaving mercuric sulphide insoluble. The extract evaporated 
to dryness and then to SOa fumes, after addition of sulphuric acid, expels nitric 
acid. Upon adding water to the residue and boiling with a Uttle additional 
sulphuric acid the sulphates of bismuth, copper and cadmium are dissolved out, 
lead sulphate remaining as a white residue. 

Lead may be further confirmed by dissolving the sulphate in ammonium 
acetate (barium sulphate is very slightly soluble,) and precipitating the yellow 
chromate, PbCrO*, by addition of potassium dichromate solution. 

ESTIMATION 

The determination of lead is required in valuation of its ores — galena, 
PbS; anglesite PbS04; cerussite, PbCOa; krokoite, PbCrOi; pyromorphite, 
SPbjPjOs'PbClj. It is determined in lead mattes; certain slags; drosses from 
hard lead; cupel bottoms; skimmings; lead insecticides (arsenate of lead); 
paint pigments such as white lead, red lead, yellow and red chromates, etc. 
It is determined in alloys such as solder, type metal, bell metal, etc. The esti- 
mation is necessary in the complete analysis of a large number of ores, especially 
in minerals of antimony and arsenic. Traces of lead are determined in certain 
food products where its presence Ls undesirable. 

Preparation and Solution of the Sample 

In dissolving lead, its alloys, or ores the following facts will be recalled. 
Hot, dilute nitric acid is the best solvent of the metal. Lead nitrate is insol- 

* Lead precipitates best from solutions containiD^ 1 cc. of concentrated free hydro- 
chloric acid (sp.gr. 1.19) for each 100 cc. of solution. The sulphide is appreciably 
soluble if the acidity is increased to 3 cc. HCl per 100 of solution. 

233 



234 LEAD 

uble in concentrated nitric acid, but dissolves readily upon dilution with water. 
The metal is insoluble in dilute sulphuric acid, but dissolves in the hot, concen- 
trated acid. Although not soluble in dilute hydrochloric, it dissolves in the 
hot, concentrated acid, especially in presence of the halogens chlorine, bromine 
and iodine. The metal is soluble in glacial acetic acid. The salts are soluble in 
hot, dilute nitric acid. In dissolving sulphide ores it should be kept in mind that 
strong nitric acid will form some lead sulphate which will be precipitated upon 
dilution of the solution. Oxidation is less apt to occiu* with the dilute acid. 
Silicates and slags require fusion with sodium carbonate and potassium car- 
bonate. The cooled mass may then be extracted with hot water to remove silica 
and the residue containing the carbonates of the heavy metals dissolved in 
dilute nitric acid. Lead salts are soluble in ammoniiun acecate. 

Ores, Minerals of Lead, etc. One gram of ore if rich in lead (galena), 
or more if the lead content is low, is placed in a platinum dish and 40 to 50 cc. 
of a mixture of one part concentrated sulphuric acid (sp.gr. 1.84) and three parts 
of concentrated nitric acid (sp.gr. 1.42) added. The covered dish is heated 
gently until the violent action has ceased, the cover is then rinsed off and 10 
to 15 cc. of hydrofluoric acid, HF, added. The mixture is evaporated to S0| 
fumes (hood) , but not to dryness, and cooled. The concentrate is diluted with 100 
cc. of distilled water and digested on the steam bath until the salts are in solution. 
The insoluble lead sulphate is filtered and washed with 10% sulphiu*ic acid solu- 
tion and finally with 50% alcohol. 

It may be advisable, in certain cases, to open up the ore with nitric acid 
or aqua regia, followed by sulphuric acid and hydrofluoric acid. 

Iron Pjrrites and Ores with Large Amounts of Impurities with Small 
Amounts of Lead. Ten grams of the sample or more, if lead is present in 
very small amounts (less than 0.1%), are taken for analysis, and 50 cc. of a 
mixture of potassium bromide and bromine solution added (75 grams of KBr 
dissolved in 400 cc. of water and 50 cc. of bromine added). After ten to 
fifteen minutes about 50 cc. of concentrated nitric acid are added and after the 
violent reaction has ceased 25 to 30 cc. of concentrated hydrochloric acid and the 
solution is evaporated on the hot plate to near dryness. Fifty cc. of C.P. (lead 
free) concentrated sulphuric acid is now added and the sample taken to fumes 
of SOi on a sand bath. After cooHng, the concentrate is diluted to 500 cc. 
with water, about 5 cc. of strong sulphuric acid added, the solution heated to 
boiling and cooled. The precipitate is filtered by decantation onto a fine-grained 
filter (quality of an S. & S. 590 or B. & A. grade A), the residue boiled with more 
water containing H2SO4 and again decanted. This is repeated until all the iron 
sulphate is removed. (The filtrates should be kept several hours to see whether 
any of the lead has passed through the paper in a colloidal condition.) The 
precipitate is finally poured on the filter and washed with 2% H2SO4. Impure 
residues are extracted for lead with ammonium acetate. 

Solution of Lead Alloys. As a rule these are lx)st decomposed by treating 
0.5 to 1.0 gram of the material, or more as the case may require, with a hot 
solution of nitric acid, 1:1, and evaporating the solution to low bulk, but not 
to dryness. Hot water is now added and the material boiled and the soluble 
portion filtered off". The insoluble material is digested with concentrated hydro- 
chloric acid to which a little bromine has been added. Boiling the mixture 
will generally eff'ect solution. (It must be remembered that lead chloride is 
diflScultly soluble in cold dilute solutions.) The lead is converted to PbSO* 



LEAD 235 

by addition of sulphuric acid and taking to S0» fumes as in case of ores. The 
purification of the impure sulphate will be given later. 

Lead may be precipitated as the chloride in the presence of a large excess of 
absolute alcohol and filtered free practically 'rom impurities. 

Brass and bronze may be dissolved in hot dilute nitric acid, 1:1. Bearing 
metal is best treated with a mixture of hydrochloric acid 5 parts and nitric acid 
1 part. 

SEPARATIONS 

Separation of Lead as Sulphate. Lead is most frequently separated from 
other metals by precipi ation as sulphate, PbSO*, according to the details given 
under " Preparation and Solution of the Sample." In the presence of much bis- 
muth or iron it is necessary to wash the precipitate with a 10% sulphuric acid 
solution to ke p the bismuth in solution and to prevent the formation of the dif- 
ficultly soluble basic ferric sulphate. In absence of appreciable amounts of 
these elements the lead sulphate is more completely separated by adding to the 
dilute sulphiuic acid solution an equal volume of alcohol, filtering and washing the 
residue with 50% alcohol. 

Separation of Lead from Barium. In the analysis of minerals containing 
barium, the insoluble sulphate, BaS04, will be precipitated with lead. Since 
barium sulphate is slightly soluble in ammonium acetate it will contaminate the 
lead in the subsequent extraction by this reagent. The presence, however, of a 
little sulphuric acid, renders this solubility practically neglig ble. The sulphuric 
acid should not exceed 1-2% in the ammonium acetate reagent as lead sulphate 
will precipitate if sufficient sulphuric acid is added to the acetate extract. (Lead 
sulphate is precipitated almost completely if the acetate solution contains 10% 
sulphuric acid.) 

Lead may be separated from barium sulphate by digesting the mixed sulphates 
with ammonium carbonate solution, whereby the lead sulphate is transposed to lead 
carbonate and ammonium sulphate, while barium sulphate is not changed. The 
soluble ammonium sulphate may be washed out with ammonium solution followed 
by water. Since lead carbonate is s'ightly soluble in the ammonium salt, the 
filtrate is treated with hydrogen sulphide and the dissolved lead recovered as 
PbS. The residue containing lead carbonate and barium sulphate is treated with 
dilute nitric or acetic acid. Lead passes into solution, while barium sulphate 
remains insoluble. 

Extraction of Lead from the Impure Sulphate by Ammonium Acetate. The 
filter containing the impure sulphate, obtained by one of the procedures for 
solution of the sample, is placed in a casserole and extracted with about 50 cc. 
of hot, slightly ammoniacal ammonium acetate, the stronger the acetate the 
better. The clear liquid is decanted through a filter and the extraction repeated 
imtil the residue is free from lead (i.e., no test is obtained for lead with 
KaCraOT). A very effective method of extraction is by adding solid ammonium 
acetate directly to the sample on a filter and pouring over it a hot solution of 
ammonium acetate. The filtrate containing the pure lead acetate solution may 
now be examined by one of the following procedures. 

Lead sulphate containing arsenic should be dissolved in ammonium acetate, 
the extract made alkaline and lead precipitated as PbS. Arsenic remains in solu- 
tion. 



236 LEAD 

The isolation of minute quantities of lead from large amounts of other sub- 
stances is described under "Gravimetric Methods for Traces of Lead." 



GRAVIMETRIC METHODS 

Determination of Lead as the Sulphate, PbS04 

Procedure. The sample having been dissolved according to a method out- 
lined, the lead precipitated as PbSO« by addition of an excess of sulphuric acid, and 
taking to SO3 fumes, the lead sulphate is filtered off, upon cooling and diluting' 
the sample. The PbSOi is washed with water containing 10% HjSO* until free 
from soluble impurities. If insoluble sulphates or silica are present the lead must 
be purified. If such impurities are known to be absent (alloys), the sulphate 
may be filtered directly onto an asbestos mat in a tared Gooch crucible, dried, 
then ignited to dull red heat, cooled and finally weighed as PbSOi. In the analysis 
of ores, however, it is generally advisable to purify the J^ulphate. 

Purification of Lead Sulphate. Details of the procedure have been given 
under Separations — Extraction of Lead from the Impure Sulphate. The lead 
sulphate having been brought into solution by extraction with strong ammonium 
acetate solution, the excess acetic acid is volatilized by evaporation, the residue 
cooled and diluted with water. An excess of sulphuric acid is added and the 
precipitated sulphate is filtered off, washed with dilute sulphuric acid and 
alcohol, dried at alx)ut 110° C, or if preferred by ignition at dull red heat, and 
weighed. 

PbSO4X0.6831 =Pb. Pb multiplied by 100 and divided by weight of sample 
taken equals per cent. 

Notes. Lead sulphate may be precipitated from ammoniimi acetate solution 
by adding sulphuric acid until the solution contains approximately 10% H2SO4. 

An acetate extraction may not be neoessarv, as is generally the case in the analysis 
of alloys. In analysis 01 ores, however, PbSOi may be contaminated by sulphates of 
the alkahne earths and by silica. The difficultly soluble oxides of iron and alumina 
may also be present. 

If arsenic is in the sulphate it will pass into the filtrate with the lead. 



Determination of Lead as the Chromate, PbCr04 

This excellent method is applicable to a large class of materials and is of 
special value in precipitation of lead from an acetic acid solution, the method 
depending upon the insolubility of lead chromate in weak acetic acid. 

Procedure. The solution of the sample, precipitation of the lead as the 
sulphate and extraction of lead with ammonium acetate have been given in detail. 

The filtrate, containing all the lead in solution as the acetate, is acidified 
slightly with a^'otic acid and heated to boihng. Lead is precipitated by addition 
of potassium dichromate solution in excess (10 cc. of 5% KzCraO? solution are gen- 
erally sufficient >. The solution is be )iled until the yellow precipitate turns to a 
shade of orange or red.' The precipitate is allowed to settle until the super- 
natant solution is clear. (This should appear yellow with the excess of dichromate 
reagent.) The PbCM)4 is filtered onto an asbestos mat in a tared Gooch cni- 

* The yellow precipitate gives high results, since it is difficult to wash. The crys- 
talline orange or red compound may be quickly filtered and washed. 



LEAD 237 

cible, washed with water, dried in an oven at about 1 10® C. and the cooled 
compoiind weired as PbCrOi. 

Ph yiOO 
PbCrO4X0.641 =Pb. ,,,. : I — per cent Pb. 

Wt. of sample 

Notes. Impurities, such as iron, copper, cadmium, etc., in the acetate solution of 
lead seriously interfere in the chromite precipitation. These should be leached out 
with water containing a little sulphuric acid before extracting the lead sulphate with 
anmionium acetate. See remarks under section on Traces of Lead. 

If a standard solution of potassium dichromate is used in the precipitation 
of lead the excess of the reagent, upon filtering of the precipitate, may bd titrated 
and the lead determined volumetrically. A known amoimt of dichromate solution 
(added from a burette) sufficient to precipitate all the lead and about one-third of the 
volume in excess is added to the hot solution. After boiling about two minutes the 

grecipitate is filtered off quickly and washed several times with hot water. The 
Itrate, or an aUquot part of it, is made acid with 5 cc. concentrated sulphuric acid 
and titrated with standard ferrous sulphate at about 60° C, using potassium ferri- 
cyanide as an outside indicator; the end-point is a blue color produced by the slight 
excess of the ferrous salt reacting with the indicator. The excess of dichromate may 
be determined by adding 3 to 4 grams of solid potassium iodide, KI, to the solution 
diluted to about 500 cc. with water to which 15 cc. of concentrated sulphuric acid 
has been added. The liberated iodine is titrated with standard thiosiJphate, the 
end-point being colorless, with starch solution internal indicator, changing from 
blue. Bi, Sb, Ba, Sr and Ca interfere slightly. 

One cc. N/10 K,Cr,07 =0.010355 gram Pb. One cc. N/5 K.CraOT =0.02071 gram Pb. 

Determination of Lead as the Molybdate, PbMo04 

This method is rapid and has the following advantages: 

a. The sulphation of lead is avoided. 6. The acetate extraction is elimi- 
nated, c. The precipitate may be ignited, d. The ratio of lead to its molybdate 
compound is greater than either lead to PbSOi or to PbCrOi, lessening the chance 
of error through weighing. 

Cobalt, calcium, strontium and barium have little effect in presence of 
ammonium acetate. In absence of this salt they interfere slightly. 

Procedure. The ore or alloy is decomposed with nitric acid or aqua regia 
as the case may require. (Silica if present is eliminated by taking to dryness, 
dehydrating, taking up with dilute nitric acid and filtering.) To the clear liquid 
ammoniimx chloride is added and then sufficient ammonium oxalate to destroy 
the excess of free nitric acid. 

Lead is now precipitated by adding 20 to 30 cc. of ammonium molybdate 
(4 grams per liter+acetic acid) stirring the mixture during the addition. After 
boiling for two or three minutes the precipitated lead molybdate is allowed to 
settle, then filtered through pulp, washed with small portions of hot water and 
ignited over a Bunsen burner. 

The cooled residue is weighed as PbMo04. 

PbMo04 X 0.5642 =Pb. 

NoTOS. If antimony or other members of the group are present in the original 
sample it is advisable to dissolve the residue in HCl and reprecipitate the lead with 
molybdate reagent. 

If lead is in the form of the sulphide, as may be the case in a complete analysis 
of a substance, it is decomposed with hot dilute HNO9 &nd precipitated as PbMoOi. 



238 LEAD 

Electrolytic Determination of Lead as the Peroxide, Pb02 

An electric current passed through a solution of lead containing sufficient 
free nitric acid will deposit all the lead on the anode as lead peroxide. The method 
is excellent for analysis of lead alloys. 

Procedure. The sample containing not over 0.5 gram lead is brought into 
solution by heating with dilute nitric acid, 1:1. The solution is washed into 
a large platinum dish with unpolished inner surface. Twenty to 25 cc. concen- 
trated nitric acid (sp.gr. 1.4) are added and the solution diluted to about 150 cc. 

The sample is electrolyzed in the cold with 0.5 to 1 ampere current and 2 to 
2.5 volts, the platinum dish forming the anode of the circuit, a spiral platinum 
wire or a platinum crucible dipped into the solution being the cathode. Three 
hours are generally sufficient for the deposition of 0.5 gram Pb. Overnight 
is advisable, a current of 0.05 ampere being used. 

A rapid deposition of the lead may be obtained by heating the solution to 
60 to 65® C. and electrolyzing with a current NDioo = 1.5 to 1.7 amperes, the 
E.M.F. varying within wide limits. Stirring the solution with a rotating cathode 
aids in the rapid deposition of the PbOs. 

To ascertain whether all the lead has been removed from the solution, more 
water is added so as to cover a fresh portion of the dish with water. The elec- 
trolysis is complete if no fresh deposition of the peroxide takes place after half 
an hour. 

The water is siphoned off while more water is being added until the acid is 
removed, the ciurent is then broken, the dish emptied of water and the deposits 
dried at 180® C. and weighed as PbO?. 

The deposit of lead peroxide gently ignited forms lead oxide, PbO, a pro- 
cedure recommended by W. C. May,* confirmed by Treadwell and Hall as 
giving more accurate results than the peroxide, Pb02. 

PbO,X0.8662=Pb. 
PbOX0.9283=Pb. 

Note. The deposits of lead oxide or peroxide may be removed by dissolving off 
with warm dilute nitric acid. 

For volumetric procedure-titration of the peroxide PbOs see page 240. 



VOLUMETRIC METHODS 
Volumetric Ferrocyanide Method for the Determination of Lead 

Although the gravimetric methods for the determination of lead are con- 
sidered the more accurate, yet the volumetric procedures may be frequently used 
with advantage. The ferrocyanide method has been pronounced by Irving C. 
Bull • to be the best of the procedures in common use, the results being accurate. 

Procedure. Lead sulphate is obtained according to the method outlined 
under Preparation and Solution of the SaTnple. The lead sulphate is transferred 
to a small beaker and gently boiled with 10 to 15 cc. of a saturated solution of 
ammonium carbonate, the liquid having been added cold and brought up to 

* Am. Jour. Sri. and Arts (3) 6, 255. 
« C. N., 2253, 87, 1903. 



LEAD 239 

boiling. After cooling, the precipitate is filtered onto the original filter paper 
from which the lead sulphate was removed. The lead carbonate is washed free 
of alkali with cold water. The filter with the precipitate is dropped into a flask 
containing a hot mixture of 5 cc. of glacial acetic acid with 25 cc. of water. 
The lead carbonate is decomposed by boiling and the solution diluted to 150 cc. 

Titration. The sample warmed to 60^ C. is titrated with a standard solu- 
tion of potassium ferrocyanide, using a saturated solution of uranium acetate, 
as an outside indicator. The excess of ferrocyanide produces a brown color with 
the uranium acetate drop on the tile. 

Free ammonia must be absent, as it reacts with uranium acetate and gives 
low results. NH4OH precipitates reddish brown, gelatinous uranous hydroxide, 
U(0H)4. 

The bulk of solution to be titrated should be as near as possible to 100 cc, 
including 10 cc. of 50% acetic acid. 

One per cent potassium ferrocyanide reagent is used in the titration. This 
reagent is standardized against a known amount of lead in solution as an acetate. 

A correction of 0.8 cc. is generally necessary on account of the indicator. 
This is determined by a blank titration. 

Antimony, bismuth, barium, strontium and calcium interfere only to a very 
slight extent, the error being negligible. 

Volumetric Determination of Lead by the Molybdate Method ^ 

Lead is precipitated as molybdate from an acetic acid solution by a standard 
molybdic solution, the termination of the reaction being recognized by the 
yellow color produced by the excess of molybdic reagent when a drop of the 
mixtiire comes in contact with a drop of tannin solution used as an outside 
indicator. 

Special Reagents Required. Ammonium Molybdate Reagent. 4.75 grams 
of the salt are dissolved in water and made up to 1 liter. One cc. with a half 
gram sample is equal approximately to 1% Pb. 

Standardization of Ammonium Molybdate Reagent. 0.293 gram pure 
lead sulphate, PbSO*, equivalent to 0.2 gram Pb, is dissolved in 50 cc. of a sat- 
lu^ated solution of ammonium acetate, a piece of litmus paper is thrown in and 
a few drops of acetic acid added to acid reaction. The solution is made up to 
200 cc. and is titrated as directed below in the procedure for lead. 

The lead value per cc. = '- : — -z gram Pb. 

cc. reagent requu'ed 

Note. In place of PbSO* pure lead foil may be taken. 0.2 gram of the foil dis- 
solved in 10-15 cc. hot nitric acid 1 : 1 and converted to the siuphate by taking to 
fumes with 20 cc. 1 : 1 H1SO4. 

Tannin Indicator. Freshly prepared tannin solution containing 0.1 gram 
tannin per 20 cc. of water. 

Procedure. 0.5 gram of the ore is dissolved by gently heating with 10 cc. of 
strong hydrochloric acid followed by 5 cc. of nitric acid and additional hydro- 
chloric acid if necessary. Five to 10 cc. of concentrated sulphuric acid are added 
and the solution evaporated to SOs fumes over a free flame. About 25 cc. of 

^Method of D. H. H. Alexander, modified by Low. 



240 LEAD 

water are added to the cooled solution and the liquid boiled for ten to fifteen 
minutes to dissolve the anhydrous ferric sulphate that may be present. 

Upon cooling, the precipitated PbSOi with any impurities it may contain 
(SiOj, CaSOi, BaSOi, etc.) is filtered off and washed with cold dilute sulphuric 
acid (1 : 10). 

Purification of the Lead Precipitate^ in Presence of Calciunii Iron, etc. 
The precipitate is rinsed into the original flask and about 5 grams of pure 
ammonium chloride and 1 cc. of concentrated hydrochloric acid added. The 
solution with the precipitates is boiled until only the silica remains imdissolved. 
The free acid is just neutralized with ammonia and the lead precipitated as 
PbS by addition of ammonium sulphide. The precipitate is filtered and washed 
free of calcium. If iron is present it must be removed by redissolving the pre- 
cipitate in 5 cc. of dilute sulphuric acid and again precipitating the lead as PbS 
by addition of sufficient hydrogen sulphide water or passing the gas into the 
acid solution. The lead sulphide is now decomposed by boiling with 5 cc. of 
concentrated hydrochloric acid for several minutes and then adding 3 or 4 drops 
of nitric acid to remove the last traces of HjS. 

The free acid in the solution is neutralized with ammonium hydroxide 
(litmus indicator), and then made slightly acid by addition of glacial acetic acid. 
The mixture is diluted to 200 cc. with hot water. 

Titration. To about two-thirds of the sample, the standard-ammonium 
molybdate is added from a burette until a drop of the solution, brought into 
contact with a drop of the tannin indicator upon a white porcelain tile or par- 
affined paper, gives a brown or yellow color. Some more of the lead solution is 
added to this titrated sample and the operation is repeated. By keeping a por- 
tion of the sample in reserve it is possible to obtain the exact end-point and avoid 
overrunning, as would be apt to occur if the entire sample were taken at one 
time. 

Cc. molybdate reagent multiplied by value in terms of Pb divided by wt. of 
sample =Pb. 

Blank. Deduction of 0.7 to 1 cc. is frequently necessary. The exact amount 
may be determined by taking the same amount of reagents as are present in 
the sample, without the lead, and titrating with ammonium molybdate, as above, 
on a boiled sample. 

Interferences: Antimony, bismuth, barium, strontium and calcium have a 
slight effect on the results. 

The lead is obtained in solution in a comparatively pure form by extraction 
of the sulphate with ammonium acetate. The more tedious method of isolation 
as directed in the procedure may not be necessary. 

Reduction and Titration of Lead Peroxide Deposited Electroljrtically 

The electrolytic deposition of lead as the peroxide, PbOj, has been given on page 
238. To avoid error that may result from imperfect drving, a volumetric procedure is 
suggested. The peroxide is dissolved from the electrode with a hot mixture of 25 cc. 
N/5 oxalic acid and 10 cc. nitric acid (sp.gr. 1.2). The excess of oxalic acid is titrated 
hot with N/10 potassium permanganate. 

1 cc. N/5 oxalic acid is equivalent to 0.02071 gram lead. 



LEAD 241 



DETERMINATION OF SMALL AMOUNTS OF LEAD 

The determination of minute quantities of lead is required in baking powders 
canned goods and like products in which small amounts of lead are objection- 
able. Traces of lead ranging from 5 to 100 parts per million (0.0005 to 0.01 % Pb) 
are best determined colorimetrically on 0.5 to 1 gram samples; larger amounts 
of lead should be determined gravimetrically. 

Gravimetric Methods for Determining Traces of Lead 

The determination of extremely small amounts of lead cannot be accom- 
plished by the usual methods of precipitation, as the lead compounds remain in 
solution in a colloidal state. The addition, however, of certain substances, 
which form amorphous precipitates with the reagents used for throwing out lead 
causes the removal of lead from the solution by occlusion. For example, the 
addition of a sufficient quantity of a soluble salt of mercury, copper, or arsenic 
to a solution containing a trace of lead, and then saturating the solution with 
HjS, will caiise the complete removal of lead from the solution. Iron and 
alumina thrown out of the solution as hydroxides will carry down small amounts 
of lead, and completely remove it from the solution, if they are present in 
sufficient quantity. Lead may be extracted from finely pulverized substances 
by means of hot anmionium acetate and precipitated from the extract as lead 
sulphide. Advantage may be taken of these facts in determining traces of 
lead in presence of large amounts of other substances. 

Amount of the Sample. It is advisable to have the final isolated lead 
compoimd over 0.01 gram in weight, hence, in a sample containing 10 parts of 
lead per million, 800 to 1000 grams of the material should be taken, since a kilo- 
gram of the material would contain 0.01 gram, Pb or 0.0156 gram PbCrO*, 
or 0.01464 gram PbS04, or 0.0177-}- gram PbMoO*. Large samples should be 
divided into several portions of 100 to 250 grams each, the lead isolated in each, 
and the final extracts, containing the lead, combined. For the given amount of 
occluding agent, stated in the procedure, the treated portion should contain not 
over 0.01 gram lead. 

I. — Extraction of Lead with Ammonium Acetate and 

Subsequent Precipitation 

It is frequently desirable to extract the lead from the mass of material and 
precipitate it from the liquor thus obtained. The procedure worked out by the 
writer is applicable to determining traces of lead in aluminum salts, but with 
modifications may be applied to a wide range of substances. 

Extraction of Lead. The desired weight of finely powdered substance, in 100- 
gram portions, is placed in 6-inch porcelain c asseroles (1000 cc. capacity) . To each 
portion are added, with vigorous stirring, 500 cc. of leid-free^ boiling hot jnmij- 
nium r.cetate solution (33%).^ The reaction is apt to be energetic, so that 

*The reagent must be boiling, when added, to obtain best results. Experiments 
have shown that considerable alumina and iron dissolve if the proportion of the reagent 
falls much below 5 cc. of 33 '-o acetate per gram of sample. With twice this amount of 
reagent the extract is free from iron and alumina. Small amounts of alumina and iron, 
however, do not interfere in the lead determination. 



242 LEAD 

care must be exercised to avoid boiling over. The residue from aluminum salts 
is crystalline and may be separated from the extract very readily by filfering 
through two filter papers in a large Biichner funnel and applying sue ion.» 
The residue is tamped down to squeeze out the adhering extract and washed 
with 100 cc. more of hot ammonium acetate followed by 100 to 200 cc. of hot 
water, again tamped down and sucked as dry as possible. The lead extracts 
are now combined and lead precipitated as sulphide. 

The reagent is made by dissolving one part cf lead-free ammonium acetate 
in two parts of diUilled water. The purity of the reagent should be tested. 

Precipitation of Small Amounts of Lead. To the solution containing lead is 
added 2-3 cc. of a 10% copper sulphate or cadmium sulphate reagent. Hydrogen 
sulphide is passed into the liquor until it is saturated. The coppe " or cadmium sul- 
phide assists the settling of lead sulphide. Gently warming on the steam bath for 
half an hour coagulates the precipitate and facilitates settling. The liquor is 
decanted through a double filter in a small Btichner funnel and the residue washed 
onto the filter with water saturated with H2S gas. 

The precipitate is washed several times with ammonium sulphide to remove 
sulphides of the arsenic group and the residue then dissolved in a hot mixture of 
hydrochloric and nitric acids ( 1 part HCl . 5 parts HNO3 and 15 parts H2O) . Ten cc. 
of strong sulphuric acid are added to the solution, and the mixture is evaporated 
to S0» fumes but not to dryness. The residue is taken up with 100-125 cc. of water 
containing 2 cc. of sulphuric acid and boiled to dissolve the soluble salts of iron, 
alumina, copper, etc. After cooling, one-third the volume of 95% alcohol is 
added (30-40 cc), the lead sulphate allowed to settle for an hour or more, then 
filtered and washed several times with 30% alcohol. The residue is extracted 
with hot ammonium acetate and lead chromate precipitated from the filtrate,' 
made slightly acid with acetic acid, by adding 10 cc. of potassium dichromate 
reagent and boiling, according to the standard procedure. (Page 236.) 

PbCrO4X0.641 =Pb. 
n. — Precipitation of Lead by Occlusion with Iron Hydroxide 

Wilkie found • that ferric hydroxide has the property of occluding lead, five 
parts of Fe(OH)s removing one part of lead from solution. Advantage is taken 
of this property of iron hydroxide in precipitating small amounts of lead.* 

Procedure. The required amount of material is weighed out in 50-gram lots 
and brushed into No. 8 beakers. If the material contains organic matter, it is 
treated with 200-cc. portions of concentrated hydrochloric acid, the mixture heated 
just below the boiling-point of HCl solution, and potassium chlorate added, a few 
crystals at a time, until the organic matter is decomposed (hood). If the material 
dissolves in water, the water solution is treated with 5 cc. of concentrated hydro- 
chloric acid and a few crystals of potassium chlorate and the liquor boiled. 

* 200 to 300 grams of material may be handled in a 6-inch Biichner funnel. 

' Should lead chromate fail to precipitate, the solution should l)e treated with 
HtS to complete saturation, the sulphide collected on a filter, then dissolved in 
acid and the procedure described above repeated. If the solution still remains clear, 
the absence 01 lead is confirmed. 

» J. M. Wilkie C. N., 2597, 117, 1909. 

* Occlusion of lead by zinc sulphide, precipitated by HjS from a formic acid solu- 
tion, is suggested; iron and alumina would not interfere. 



LEAD 243 

Addition of Ferric Iron. If sufficient iron is not already present, ferric 
chloride is added in such quantity that the iron content of the sample will be from 
twenty to fifty times that of the lead (larger amounts of iron will do no harm) 
present in the solution. Five to 10 cc. of concentrated nitric acid are added 
and the sample boiled for ten to fifteen minutes. 

Precipitation of Iron and Lead. If alumina is present, iron is precipi- 
tated by addition of a large excess of potassium hydroxide, the aliunina going 
into solution as potassium aluminate. In absence of alumina, ammonia may 
be used to precipitate the ferric hydroxide. Lead is completely occluded by the 
precipitate and carried down. The solution is filtered hot through Baker and 
Adamson's fast filters, threefold. The filtering must be rapid and the liquid kept 
hot to prevent clogging of the filters. 

Separation of Lead from Iron. The precipitate is dissolved in hot hydro- 
chloric acid (free from lead). The solutions are combined, if several portions 
of the sample are taken. Concentrated sulphuric acid is added and the sample 
evaporated to small volume and heated until the white sulphuric acid fumes 
appear. The usual procedure is now followed for separation of the lead sulphate, 
acetate extraction of lead and final precipitation of lead chromate. 

PbCrO4X0.641 =Pb. 

Note. In place of using alcohol to decrease the solubility of lead sulphate, many 
prefer to add sulphuric acid so that the acidity of the solution will be 2-10% free 

111. Modification of Seeker-Clayton Method for Traces of Lead 

in Baking Powder 

One hundred grams of baking powder are treated with 25 cc. of water followed 
by 75 cc. of strong hydrochloric acid added in small portions to avoid excess frothing. 
The mixture is heated until the starch has decomposed (iodine test gives blue color 
with starch), the solution becoming clear and turning yellow. The free acid is 
neutralized and when the solution is cold, 400 cc. of lead-free ammonium citrate, 
saturated with HaS, are added. Additional H^S is passed into the slightly alkaline 
solution, the sulphides of iron and lead allowed to settle, the clear supernatant 
liquor decanted off, the sulphides collected on a filter and washed. The precipitate 
is dissolved in nitric acid, lead separated as a sulphate, extracted with acetate and 
precipitated as dichromate according to the procedure recommended under the 
acetate extraction method I.^ 



COLORIMETRIC ESTIMATION OF SMALL AMOUNTS 

OF LEAD 

Introduction. Estimation of small amounts of lead by the intensity of the 
brown coloration produced by the sulphide in colloidal solution was first proposed 
by Pelouze.' The procedure was modified by Warington » and by Wilkie * 

1 See Referee's modification Jour. Assoc. Off. Ag. Chemists, 1, 3, 512 (Nov., 1915.) 
« T. J. Pelouze, Ann. Chim. Phys., 8, 79-108, 1841. 
» R. Warington, Jour. Soc. Chem. Ind., 12, 97, 1893. 
* J. M. Wilkie, Jour. Soc. Chem. Ind., 28, 636, 1909. 



244 LEAD 

to overcome the color produced by accompanying impurities, among these, of 
iron, which is ahnost invariably associated with lead. The method is useful in 
determining traces of lead in drinking water, in food products, baking powders, 
canned goods, phosphates, alums, acids such as sulphuric, hydrochloric, citric, 
tartaric and the like. By this procedure on a gram sample one part of lead per 
million may be detected and as high as 50 parts may be estimated. For larger 
amounts of lead, a smaller sample must be taken. Nickel, arsenic, antimony, 
silver, zinc, tin, iron, and alumina, present in amounts such as commonly occur 
in these materials, do not interfere.* 

In order to obtain accurate results it is necessary to have the solutions \mder 
comparison possess the same general character. ** It must be remembered that 
the tint depends to a large extent on the size of the colloidal particles of lead, 
which in turn depend upon the nature of the salts in the solution and upon the 
way that the solution has been prepared." • Vigorous agitation, salts of the 
alkalies and alkaline earths tend to coagulate the colloidal sulphide. 

Reagents Required. Standard Lead Solution, A convenient solution 
may be made by dissolving 0.1831 gram of lead acetate, Pb(C2H|Oi)t'3H20 
in 100 cc. of water, clearing any cloudiness with a few drops of acetic acid and 
diluting to 1000 cc. If 10 cc. of this solution is diluted to 1000 cc. each cc. 
will contain an equivalent of 0.000001 gram Pb. 

Harcourt suggests a permanent standard made by mixing ferric, copper and 
cobalt salts.' For example 12 grams of FeCU together with 8 grams of CuClj 
and 4 grams of Co(N03)a are dissolved in water, 400 cc. of hydrochloric acid 
added and the solution diluted to 4000 cc. 150 cc. of this solution together 
with 115 cc. of hydrochloric acid (1 : 2) diluted to 2000 cc. will give a 
shade comparable to that produced by the standard lead solution above, when 
treated with the sulphide reagent. The exact value per cc. may be obtained by 
comparison with the lead standard. 

Alkaline Tartrate Solution. Twenty-five grams of C.P. sodium potassiiun 
tartrate, NaKC4H40« '41120, is dissolved in 50 cc. of water. A little ammonia is 
added and then sodium sulphide solution. After settling some time the reagent 
is filtered. The filtrate is acidified with hydrochloric acid, boiled free of H2S and 
again made ammoniacal and diluted to 100 cc. 

Ammonium Citrate Solution, Ammonium citrate solution is prepared in 
the same way as the tartrate solution above, 25 grams of the salt being dissolved 
in 50 cc. of water. 

Potassium Cyanide, Ten per cent solution. The salt should be lead-free. 

Sodium Sulphide. Ten per cent solution, made from colorless crystals. 
Sodium sulphide may be made by saturating a strong solution of sodium hydroxide 
with hydrogen sulphide gas, and then adding an equal volume of the sodium 
hydroxide. The solution is diluted to required volume, allowed to stand several 
days, and filtered. 

Sodium metabisulphite. Solid salt of NaSiOi. * 

Apparatus. The color comparison may be made in Nessler tubes, or in a 
colorimeter. The Campbell and Hurley modification of the Kennicott-Sargent 

1 Ni up to 0.1%, As up to 0.2%, Zn 0.2%, Sb 0.05%, Cu 0.25%, Fe 1.0%, Al 10%, 
Sn up to 1.4% do not interfere. 

« J. W. Mellor, " A Treatise on Quantitative Inorganic Analysis." 
» A. G. V. Harcourt, Jour. Chem. Soc, 97, 841, 1900. 
* Recommended by W. S. Allen for reduction of iron. 



LEAD 



245 



colorimeter is excellent for this purpose,' Fig. 43. The cotorimeter is simple in 
construction and operation. 

The tubes for holding the solutions to be compared are those of one of the 
well-known colorimeters, in which the unknown solution ia placed in the left-hand 
tube while the color is matched by 
raising or lowering' the level of a 
standard solution in the right-hand 
tube by means of a glass plunger , 
woridug in an attached reservoir. 

The accompanjing diagram 
sbowB the essential features of con- 
struction of the colorimeter em- 
ployed in the tests described below. 
The unknown solution is placed in 
the left-hand tube A, which is 19 cm. 
long, 3 cm. in diameter, and gradu- 
ated for 15 cm. The standard solu- 
tion is placed in the right-hand 
tube B, which is the same size as 
A, the graduated portion being 
divided into 100 divLsiona of 1 .5 mm. 
each. The tube B is permanently 
connected by a glass tube with the 
reservoir C in which the glass 
plunger D works, so that the level 
of the liquid in B can be readily 
controlled by raising or lowering the 
plunger. ^ the tube £ and reser- 
voir C are made in one piece, the 
liquid used for the standard solution 
cornea in contact with glass only, 
thus preventing any possibility of 




FiQ. 43. — Hurley's Colorimeter. 



chemical change due to contact with the container. The plunger is provided 
with a rubber collar E, so placed as to prevent the plunger from accidentally 
striking and breaking the bottom of the reservoir. The tubes A and B, with 
the connecting reservoir, rest on wooden supports, the one under A and B being 
provided with holes for the passage of the light, and are held in position by 
spring clips F F. This arrangement allows the glass parts to be readily removed 
for cleaning and filling. The light for illuminating the solution is reflected 
upward through the tubes A and B by means of the adjustable mirror G. The 
best results are obtained by facing the colorimeter toward a north window in order 
to get reflected skylight through the tubes, care being taken to avoid light reflected 
from adjacent objects. The black wooden back of the colorimeter serves the 
double purpose of a support for the parts of the instrument and of a screen, as 
it ia interposed between the color tubes and the source of light. 

The light, passing upward through the tubes A and B, impinges on the two 

mirrors H and I cemented to brass plates sliding in grooves cut pt an angle of 

45° in the sides of the wooden box J. This box is supplied with a loosely-fitting 

cover, thus allowii^ easy access for the purpose of removing and cleaning the 

■Jour. Am. Chem. Soc, 88, 1112, July, 1911, 



246 LEAD 

mirrors. The mirror H is cut vertically and cemented in such a position as to 
reflect one-half of the circular field of light coming through the tube A, The 
light passing upward through B is reflected horizontally by the mirror /, through 
a hole in the brass plate supporting the mirror H, One-half of the circular field 
of light from the tube B is cut off by the mJrror H, the vertical edge of which 
acts as a dividing line between the two halves of the circular field. The image 
of one-half of the tube B is then observed in juxtaposition to the opposite half 
of the image of the tube A, 

The juxtaposed images are observed through a tube K^ 2.5 cm. in diameter 
and 16 cm. long, lined with black felt and provided with an eye-piece having a 
hole 1 .5 mm. in diameter. At the point M in the tube K is placed a diaphragm 
having an aperture 8 mm. in diameter. All parts inside the box J except the 
mirrors are painted black so that no light except that coming through the tubes 
A and B passes through the tube K. By having the apertures in the eye-piece 
and diaphragm properly proportioned only the image of the bottoms of the 
tubes A and B can be seen, thus preventing interference of light reflected from 
the vertical sides of the tubes A and B. 

A person looking through the eye-piece observes a single circular field divided 
vertically by an almost imperceptible line when the two solutions are of the same 
intensity. By manipulating the plunger D, the level of the liquid in B can be 
easily raised or lowered, thus causing the right half of the image to assume a 
darker or lighter shade at will. In matching colors with an ascending colunr.n 
in B, that is, gradually deepening the color of the right half of the field, the 
usual tendency is to stop a little below the true reading while in a comparison 
with a descending column the opposite is the case. 

Procedure. If lead is between 10 to 50 parts per million a 1-gram sample 
is taken. If it is above or below these extremes the amount of sample is regu- 
lated accordingly. In materials containing organic matter it is not advisable 
to take more than a 1-gram sample. 

Substances containing organic matter, such as starch in baking powder, 
should be decomposed by fusion with sodium peroxide, sodium or potassium 
sulphate containing a few drops of sulphuric acid. A Kjeldahl digestion with 
concentrated sulphuric acid and potassium bisulphate may occasionally be 
advisable. Sulphuric acid discolored by organic matter should be mixed with 
4 to 5 grams of potassium bisulphate, taken to fumes ind then diluted with 
water. The material may be extracted with ammoniim acetate and lead 
determined in the extract. See notes. 

To the solution containing the sample are added 10 cc. of tartrate solution 
(or 20 cc. of citrate solution with phosphates of lime, etc.), 10 cc. of hydrochloric 
acid and the mixture brought to boiling. Small amounts of ferric iron are now 
reduced by adding 0.5 gram sodium metabisulphite. Sufficient ammonium 
hydroxide is added to neutralize the free acid and 5 cc. in excess; then 3 cc. potas- 
sium cyanide (to repress any copper color that may be present to reduce higher 
oxides), and the mixture heated until the solution becomes colorless. The entire 
solution or an aliquot portion is placed in the comparison cylinder, and diluted 
to nearly 100 cc. If the Kennicott-Sargcnt apparatus is used the standard color 
solution is forced into tlic adjacent cylinder, until the color in tliis cylinder matches 
the one containing tlie sample. The number of cc. of the standard is noted. This 
blank is due to the slight color that the solutions of the samples invariably have. 
Four drops of the sulphide reagent are added to the sample and this is mixed 



LEAD 247 

by means of & plunger, avoiding any more agitation than is absolutely necessary 
to make the solution homogeneous. After one minute the compariaun is iigain 
made, the colored standard being forced into the cylinder until its color matches 
the sample. It is advisable to take several readings with ascending and descend- 
ing column of standard reagent, takin;; the average as the true reading. 

Calculation. Suppose the standard =0.00000) gram Pb per ce., blank =5 cc, 
total reading -22 cc, one gram of sample l)cing taken for analj-sis. Then 
22-5-17 cc. -0.0017% Pb or 17 parts per million. 



NoTKB, Iron must be completely reduced before adding ainmoQium hydroxide 
and 7>ota.'Qiuni cyanide. 

Allm's method of reducing iron with sodium metnbisul])hite ia excellent. The salt 
mav be made by passing NO] into a saturated Holution of sodium carbonate at 
boiling temperature, until the licjuor is just acid to methyl orange. The water 
evEHMiated during the treatment is replaced during the action. NaiStOt separates 
and may be filtered off and the water removed by centrifugiDK. 

An 

FiQ. 44.^-Cooper Hewitt Mercury Light. 

The Cooper Hewitt Mercury light is excellent for colorunetric lead determina- 
tions, where an artificial light is desired. The yellow shades appear yellowish- 
green and may be matched more readily than the yellows obtained by daylight. 

The illustration. Fig. 44, shows the type of light recommended for this work. 

Ifaseporatioft from iron is desired, the lead may lie extracted with ammonium 
acetate solution. Ten grams of the powdered material are mixed with 75 cc. of 
a 33% aitmonium acetate solution ' (25 grams of the salt dissolved in 50 cc. HiO), 
the reagent being added boiling hot. The mixture is diluted to 500 cc, a portion 
filtered, and the dctcnninatiou made on an aliquot part of the total, followii^ 
the directions above. 

' The ammonium acetate should be free from lead. 



248 LEAD 

ANALYSIS OF METALLIC LEAD 

Determination of Impurities in Pig Lead— Complete Analysis ^ 

The following substances are generally estimated in the complete analysis 
of lead: silver, bismuth, copper, cadmium, arsenic, antimony, tin, iron, cobalt, 
nickel, manganese and zinc. 

Determination of Silver 

This is determined by assay of 100 grams of lead. The substance is placed 
in a 3-in. scorifier and heated in a muffle furnace until the assay " covers.*' It 
is then poured into a mould, allowed to cool and the button thus obtained again 
scorified until a final button weighing about 20 grams is obtained. This is 
cupeled and silver determined as usual. If the silver bead is large it should be 
parted for gold. 

Determination of Bismuth 

In determining bismuth three cases arise: A. The ordinary method. B, 
Procedure for determining minute amounts of bismuth. C Method in presence 
of comparatively large amounts of antimony and tin. 

A. Twenty grams of lead are dissolved in 100 cc. of hot dilute nitric acid 
(1:4). If the solution is complete, dilute ammonium hydroxide is added, drop 
by drop, until a faint opalescence is observed in the solution. If a precipitate is 
formed, this must be dissolved by addition of nitric acid and the ammonia 
treatment repeated. Now 5 cc. of dilute hydrochloric acid are added (1:9) 
and the solution diluted to 400 cc. and heated to boihng. The bismuth oxy- 
chloride is allowed to settle on the steam bath for several hours, the clear solu- 
tion is then decanted through a 7 cm. filter (S. & S. No. 589), the precipitate 
transferred to the paper and washed with hot water. (The solution is refiltered 
if cloudy.) The precipitate is dissolved with 5 cc. hot hydrochloric acid (1 : 9), 
the acid being added around the edge of the filter with a pipette. The paper 
is washed and the solution diluted to 400 cc. and brought to boiling. The 
precipitate is filtered into a weighed Gooch crucible, washed several times with 
water, then once with alcohol and finally with ether. It is dried in the oven and 
weighed as BiOCl. 

BiOClX.802=Bi. 

B. Determination of minute amounts of bismuth is made as follows: 100 
grams of lead are dissolved in 500 cc. of dilute nitric acid (1 : 4), and the cooled 
solution treated with sufficient saturated solution of sodium carbonate to pro- 
duce a heavy precipitate. After settling, then decanting off the clear solution, 
the precipitate is filtered onto a filter and drained. Without washing this is 
dissolved with the least amount of nitric acid that is required. The solution is 
then neutralized with aimnonia as before (method A), litmus paper being used 
as an indicator, and bismuth detcnnincd as directed under the first procedure. 

C. In presence of considerable amounts of antimony and tin, the bisnmth is 
precipitated as in case A, the precipitate dissolved in hot hydrochloric acid 
(1 : 2), and the solution diluted to 200 cc. The sulphides of antimony, tin, etc., 

^ Method of the National Lead Company, modified. 



LEAD 249 

are precipitated with HtS, antimony and tin dissolved out with a solution of 
potassium hydroxide and sulphide water (1 part 20% KOH to 4 parts HtS water), 
and the residue washed. This is dissolved in 20 cc. of hot nitric acid (1 : 4), and 
bismuth determined as usual in the filtrate. 

Determination of the Remaining Elements 

222.23 grams of the sample of lead are dissolved in 1100 cc. of dilute nitric 
acid (1 : 4) in a large beaker. If the solution is turbid, appreciable amounts 
of antimony and tin are indicated with possible sulphur combined as PbS04. 
In this case it is filtered into a 2000-cc. flask. If the solution is clear it is 
transferred directly to the flask. 

Residue I. May contain As, Sb, Sn, Filtrate I. Contains all the elements 
PbS04. present in the sample. 

Residue L The residue and filter is treated with 20 cc. of tartaric acid 
mixture (50 grams tartaric acid, 250 cc. of water and 250 cc. of concentrated 
hydrochloric acid). After boiling the mixture is digested on the steam bath for 
half an hour, then 50 cc. of hot water added and the solution filtered. The filter 
paper is ignited and any residue is dissolved by fusion with 1 gram of potassiiun 
hydroxide in a silver dish. The water extract of this fusion is added to the tar- 
trate solution. Now ammonia is added until the solution is alkaline and then 
hydrochloric acid until it is sUghtly acid. Hydrogen sulphide is now passed in 
to saturation, the precipitate digested on the steam bath for fifteen to twenty 
minutes and hydrogen sulphide again passed into the solution about fifteen 
minutes. The sulphides are filtered off, arsenic, antimony and tin sulphides dis- 
solved with 5 cc. (1 : 5) potassium hydroxide in 25 cc. of saturated H2S water. 
The solution is diluted to 111 cc, and 100 cc. — equivalent to 200 grams of sam- 
ple — ^preserved for subsequent analysis. This solution is marked " Extract C." 

Filtrate I* This solution, containing practically all of the material, is 
treated with 150 cc. of dilute sulphuric acid (1 : 1), and the solution made to 
volume — 2000 cc. It is now transferred to a 3-liter flask, the graduated flask 
rinsed out into the main solution with 60 cc. of water. (The PbSOi precipitate 
found to occupy space of 50 cc.) When the precipitate has settled, 1800 cc. 
are decanted off. This represents 200 grams of the sample. The solution is 
boiled down in a No. 9 porcelain evaporating dish, heating first over the free 
flame and finally on the steam bath until only a moist residue remains. Fifty 
cc. of water is added, the residue transferred to a beaker and digested for several 
hours, preferably overnight, and then filtered. 

Residue U. This may contain PbS04, Filtrate U. This may contain Cu, 
As, Sb, Sn salts. Bi, Cd, Sn, Sb, As, Fe, Co, Ni 

and Zn. 

Residue U. This is treated as has been described for residue I. The 
entire solution is added to the Extract C. The residue, consisting of PbS04, is 
rejected. 

Filtrate U. This is made neutral with ammonium hydroxide and then con- 
centrated hydrochloric acid added in such an amount that the solution will 
contain 4% free acid. (HCl sp.gr. 1.2, 4 cc. per 100 of solution.) Hydrogen 
sulphide is now passed into the hot solution imtil it is satiu*ated, the precipitate 



250 LEAD 

settled on the steam bath for half an hour and hydrogen sulphide agam passed 
in for fifteen minutes. The precipitate is filtered off and washed with H,S 
water slightly acidified with hydrochloric acid. 

Residue III. May contain CuS, Filtrate in. May contain ions of 
Bi^Sa, CdS, AS2S3, Sb,S„ SnS. Fe, Al, Co, Ni, Mn and Zn. This 

filtrate is marked " B.'* 

Residue HI. The sulphides are extracted with potassium hydroxide and 
hydrogen sulphide solution. This dissolves out arsenic, antiuiony and tin. This 
extract is combined with the extract marked " C." 

The residue remaining is marked " Residue A." 

The constituents of the sample have now been isolated in the groups. 

Residue '' A '' contains the sulphides of copper, bismuth and cadmium. 

Filtrate " B " contains such elements as do not precipitate as sulphides in 
acid solution — iron, aluminum, manganese, cobalt, nickel and zinc. 

Extract " C *' includes the elements arsenic, antimony and tin. 

Determination of Arsenic, Antimony, and Tin in Pig Lead 

The combined alkali sulphide solutions: " Extract C *' is washed into a 
beaker and acidified with 20 cc. of nitric acid and 5 cc. of hydrochloric acid. 
The solution is evaporated to dryness on the steam bath. The residue is dissolved 
in 200 cc. of water and 10 grams of oxalic acid added, together with 10 grams 
of ammonium oxalate, and the solution heated until clear. 

Hydrogen sulphide gas is now passed into the hot solution for forty-five 
minutes. 

Precipitate. AsSh SbsSs. Filtrate contains Sn. 

Arsenic. The precipitate containing arsenic and antimony is placed in a 
distilling flask, strong hydrochloric acid added and arsenic separated from 
antimony by distillation with a current of HCl gas according to the regular pro- 
cedure. If a precipitate of arsenic sulphide forms in the distillate, it is advis- 
able to precipitate the arsenic as sulphide, oxidize the compound to form 
sulphate and arsenic acid, and after reduction of the arsenic to titrate it with 
standard iodine. This oxidation may be accomplished, before distilhition with 
hydrochloric acid. For details of the procedure see chapter on Arsenic, page 33. 

Antimony is determined in the residue in the flask by titration with N/10 
potassium bromate or by the potassium iodide method. 

I. 2KBrO,4-2HCl-h3Sb20, =2KCl-h2HBr4-3Sb205. 

II. (a) SbjCl5+2KI=Sb2CU4-2KCl-hl2. 
(6) l2-h2Na2S20,=2Nal4-Na2S406. 

For details of the procedure see chapter on Antimony, pages 25 and 26. 

Determination of Copper and Cadmium in Pig Lead 

The residue " A " is taken for this analysis. If copper exceeds 0.0025% 
method I is used. If the copper percentage is below this amount the procedure 
II is followed. 



LEAD 251 

Method I. The residue is dissolved by heating with 20 cc. of nitric acid 
(1:4) and the solution filtered into a beaker. The filter is ignited and the 
residue dissolved in nitric acid (1:1) and the solution added to the first por- 
tion. The volume should not exceed 100 cc. Ammonium hydroxide is added 
until the solution is strongly ammoniacal and then 5 grams of potassium cyanide. 
Hydrogen sulphide is passed into the cold solution to saturation, and the 
solution filtered. 

Precipitate =AgS, BijSs, CdS. Filtrate =Cu in solution. 

The filtrate containing the copper is evaporated on the steam bath to a 
volume of 20 to 30 cc. in a 4-in. casserole. Now 20 cc. of sulphuric acid (1:1) 
are added (hood), and the solution evaporated until 80$ fumes are evolved. 
The cooled concentrate is diluted with water and filtered, if necessary. Three cc. 
of nitric acid are added per 100 cc. of solution and the copper deposited by 
electrolysis according to the regular procedure and weighed as metallic copper. 
For detailed method see chapter on Copper, page 155. 

The precipitate containing silver, bismuth and cadmium is dissolved in 20 cc. 
of nitric acid (1 : 4), 1 cc. of 1% sodium chloride solution is added, the solution 
digested half an hour and then filtered and the filter washed with water. 

Precipitate -AgCl, reject. FUtrate-Cd(N03)2 and Bi(N03)3. 

The filtrate is made slightly alkaline with sodium carbonate added in slight 
excess, and 5 grams of potassium cyanide are then added. After digesting on the 
steam bath for half an hour the solution is filtered and the residue washed with 
5% sodium carbonate solution. 

Precipitate contains bismuth, reject. Filtrate contains cadmium. 

The filtrate is now treated with a few cc. of ammonium sulphide and the 
yellow cadmium sulphide is filtered into a weighed Gooch crucible, then washed, 
dried and finally weighed as CdS. 

CdSX0.778=Cd. 

Method n. Small amounts of copper. The filter containing the sulphides 
is ignited in a porcelain crucible and the residue dissolved in 5 to 10 cc. of nitric 
acid (1 : 1), and the solution evaporated to pastiness. One cc. of sulphuric acid 
(1 : 1) is added together with a few drops of 10% sodium chloride solution and the 
mixture evaporated to SOa fumes, the cooled product then diluted with water 
and filtered from the lead and silver precipitates. 

Ammonia is now added to the filtrate together with 5 grams of potassium 
cyanide and CdS and BijSa are precipitated with H2S, as in case I, and filtered off. 

Precipitate— CdS and Bi^Sa. Bis- Filtrate. The solution is made acid 

muth is removed as before and in the hood with H2SO4, then taken 

cadmium sulphide again precipi- to SO3 fumes and copper determined 

tated and the compound titrated by the potassium iodide method, 
with N/10 iodine solution. 
1 cc. N/10 I =0.00562 gram Cd. 



252 LEAD 

Determination of Iron, Cobalt, Nickel, Manganese and 

Zinc in Pig Lead 

Iron and Alumina. The filtrate " B " from members of the Hydrogen 
Sulphide Group is evaporated to 100 and the iron oxidized with a few drops of 
nitric acid as usual. Iron (and alumina) hydroxide is now precipitated by addition 
of ammonia. It is advisable to dissolve this precipitate in hydrochloric acid and 
reprecipitate the iron to recover the occluded manganese and zinc. The com- 
bined filtrates are reserved for the determination of the remainder of the elements. 
The hydroxide of iron is ignited and weighed as Fe20j. If alumina is suspected, 
the residue is dissolved in hydrochloric acid and iron determined volumetrically. 
FejOj thus obtained is subtracted from the weight of the first determination, 
the difference being due to the alumina present. 

Fe,0, X 0.6994 = Fe Reciprocal factor = 1 .4298 

AUO.X 0.5303 =A1 Reciprocal factor = 1.8856 

Zinc. The filtrate from iron precipitate is made neutral with hydrochloric 
acid and then 15 drops of 2N. HCl added in excess and zinc precipitated in the 
pressure flask with HjS. (See Figs. 3 and 4 in chapter on Arsenic.) The 
sulphide of zinc is filtered off, and either ignited to the oxide ZnO and so weighed 
or determined by a volumetric procedure. See chapter on Zinc. 

ZnO X 0.8034 =Zn. 

H,S04 X 0.6665 =Zn. 

Cobalt and Nickel. These are best determined by electrolysis, being deposited 
from an ammonium sulphate solution according to the procedure described for 
these elements. 

If a separation of the elements is desired the deposit is dissolved in acid, 
nickel determined by O. Brunck's dimethylglyoxime method, and cobalt deter- 
mined by difference. 

Manganese. The solution from nickel and cobalt is taken to dryness, and the 
residue heated to expel the ammonium salts and destroy any organic matter 
present. This is taken up with a little hydrochloric acid, then 2 to 3 cc. of 
sulphuric acid added and the mixture evaporated to SOs fumes to expel the hydro- 
chloric acid. When nearly all the free acid is driven off, the moist residue, cooled, 
is treated with 50 cc. of nitric acid (1 : 3), and manganese determined in the 
solution preferably by the bismuthate method. For minute amounts of man- 
ganese the colorimetric procedure is used. See chapter on Manganese, page 267. 



MAGNESIUM 

Wilfred W. Scott 
Ms, at.wt. 24.32; sp.gr. 1.6&-1.75; m.p. 651'' '; b.p. 1120'' C.^; oxide MgO. 

DETECTION 

In the usual course of analysis magnesium is found in the filtrate from the 
precipitated carbonates of barium, calcium, and strontium. The general procedure 
for removal of the preceding groups may be found in the section on Separa- 
tions given on the following page, 254. Magnesium is precipitated as white 
magnesium ammonium phosphate, MgNH4P04, by an alkali phosphate, NaaHP04, 
NaNH4HP04, etc., in presence of ammonium chloride and free ammonia. The 
precipitate forms slowly in dilute solution. This is hastened by agitation and 
by rubbing the sides of the beaker during the stirring with a glass rod. Crystals 
soon appear on the sides of the beaker in the path of contact, and finally in 
the solution. 

Baryta or lime water added to a solution containing magnesium produces 
a white precipitate of magnesium hydroxide. 

Both the phosphate and the hydroxide of magnesium are soluble in acids. 

ESTIMATION 

The element is determined in the complete analysis of a large number of 
substances; in the analysis of ores, minerals, rocks, soils, cements, water, etc. 
The following are the more important ores in which the element occurs: Mag- 
nesite, MgCOj; dolomite, CaCOa-MgCO*; kieserite, MgS04-H20; kainite, 
MgS04KCl-6H,0; camallite, MgCUKCl-GHjO; in the silicates, enstatite, 
MgSiOa; talc, H2Mg3(SiOi)4; meerschaum, forsterite, Mg2Si04; titanate, MgTiOj; 
olivine, Mg2Si04-Fe2Si04; serpentine, H4Mg3Si204. It occurs as boracite, 
4MgB407-2MgO-MgCl2. It is found in sea-water, and in certain mineral waters. 
It occurs as a phosphate and carbonate in the vegetable and animal kingdoms, 
especially in seeds and bones. 

Preparation and Solution of the Sample 

In solution of the material it will be recalled that the metal is soluble in 
acids and is also attacked by the acid alkali carbonates. It is soluble in am- 
monium salts. The oxide, hydroxide, and the salts of magnesium are soluble 
in acids. Combined in silicates, however, the substance requires fusion with 
alkali carbonates to bring it into solution. 

General Procedure for Ores. One gram of the ore is treated with 20 cc. of 
strong hydrochloric acid and heated gently until the material is decomposed. 
If sulphides are present, 5 to 10 cc. of strong nitric acid are added and the material 
decomposed by the mixed acids. If silicates are present and the decompo- 

1 Circular 35 (2d Ed.) U. S. Bureau of Standards. 
' Van Nostrand's Chem. Annual— Olsen. 

253 



254 MAGNESIUM 

sition is not complete by the acid treatment, the insoluble material is decom- 
posed by fusion with sodium carbonate, or the entire sample may be fused with 
the alkali carbonate, the fusion is dissolved in hydrochloric acid and taken to 
dryness. Silica is dehydrated as usual by heating the residue from the evaporated 
solution. This is taken up with 50 cc. of water containing about 5 cc. strong 
hydrochloric acid, the silica filtered off and, after removal of the interfering sub- 
stances according to procedures given under the next section on Separations, 
magnesium is determined as directed in the sections on Methods. 

SEPARATIONS 

Removal of Members of the Hydrogen Sulphide Group. Copper, Lead, 
Bismuth, Cadmium, Arsenic, etc. The filtrate from silica ^ is diluted to about 
200 cc. and hydrogen sulphide gas passed in until the members of this group 
are completely precipitated. The sulphides are filtered off and washed with HjS 
water and the filtrate and wasihings concentrated by boiling. This treatment 
is seldom necessary in analysis of many silicates and carbonates in which these 
elements are absent. 

Removal of Iron, Aluminum, Manganese, Zinc, etc. The concentrated 
filtrate from the hydrogen sulphide group, or in case the treatment with hydro- 
gen sulphide was not required, the filtrate from silica, is boiled with a few cc. 
of nitric acid to oxidize the iron (solution turns yellow), about 5 cc. of concentrated 
hydrochloric acid added, and if manganese is present, 15 to 20 cc. of a saturated 
solution of bromine water, and the solution made alkaline to precipitate iron, 
aluminum, manganese. If zinc, cobalt, and nickel are present, these are best 
removed as sulphides by passing hydrogen sulphide into the ammoniacal solu- 
tion under pressure. (See Fig. 3 and Fig. 4, pages 38 and 39.) 

Separation of Magnesium from the Alkaline Earths. The alkaline 
earths are precipitated either as oxalates, recommended when considerable 
calcium is present, or as sulphates, recommended in presence of a large pro- 
portion of barium, the magnesium salts being soluble. Magnesium is pre- 
cipitated from the filtrates as a phosphate, according to directions given later. 
Details of the separation of magnesium from the alkaline earths may be found 
in the chapter on Barium, page 53. 

An excellent procedure for the separation by means of sulphuric acid is 
to evaporate the solution to dryness, concentrathig first in a porcelain dish and 
finally to dryness in a platinum dish, and then adding about 50 cc. of 80% 
alcohol and sufficient sulphuric acid to combine with the alkaline earths and 
magnesium, with slight excess. This precipitates barium, strontium, and cal- 
cium as sulphates, while the greater part of the magnesium is in solution. 
After settling, the precipitate is filtered and washed free of sulphuric acid by 
means of absolute alcohol, then with 40% alcohol to remove any magnesium 
sulphate remaining with the precipitate. Magnesium is dctennined in the 
filtrate by expelling the alcohol by evaporation, and then precipitating as mag- 
nesium anunonium phosphate according to directions given for the determination 
of this element. 

Note. Magnesium is prevented from precipitation as a hydroxide by the presence 
of ammonium salts. See note, bottom of page 8. 



1 » 



See previous paragraph. 



MAGNESIUM 255 

QRAVIMETRIC DETERMINATION OF MAGNESIUM 

Precipitation of Magnesium by a Soluble Phosphate as 
Ammonium Magnesium Phosphate 

Magnesium is determined in the filtrate from calcium oxalate by the addition 
of sodium ammonium phosphate to a hot slightly acid or neutral solution followed 
by a definite amount of anunonia. The practice of precipitating magnesium from 
a cold solution necessitates a double precipitation as the com]X)sition of the 
phosphate is considerably modified by that of the solution in which the precipi- 
tation takes place, so that it is necessary to adjust conditions by having a 
definite amount of ammonia, anmionium salts and phosphate for the approxi- 
mate amount of magnesium present.* Accurate results are obtained by pre- 
cipitation of the compound from a hot solution by the method of B. Schmitz,* 
by addition of the soluble phosphate to a slightly acid solution and then mak- 
ing anunoniacal, or that of W. Gibbs,* by precipitation of the amorphous 
magnesium hydrogen phosphate in a neutral solution and transforming the pre- 
cipitate to magnesium ammonium phosphate by addition of ammonia to the hot 
solution. Upon ignition of the precipitate, magnesium pyrophosphate (MgjPiOy) 
is formed. 

Reactions. 

A. Na,NH4P04+MgCU=2NaCl+MgNH4P04 (B. Schmitz).* 

B. NaHNH4P04+MgCU=NaCl4-NH4Cl+MgHP04 and 
MgHP04+NH,=MgNH4P04 (W. Gibbs).« 

Decomposition with Heat. 

2MgNH4P04 =2NH,-hH20+Mg,P,07. 

The following procedure gives accurate results. 

Procedure. The neutral or slightly acid solution, containing magnesium in 

presence of ammonium salts, is heated to boiling and treated, drop by drop, with 

an excess of sodium or ammonium phosphate, or microcosmic salt (10% solutions), 

stirring constantly during the addition. Then ammonium hydroxide (sp.gr. 0.96) 

is added, its volume measuring one-third that of the magnesium solution. The 

crystalline precipitate Ls allowed to cool and settle for two hours or more. The 

supernatant liquid is filtered off, the precipitate washed by decantation two or 

three times, then transfeiTed to the filter, using dilute ammonia water (2%). The 

precipitate is dried and then transferred as completely as possible to a weighed 

platinum crucible, the ash of the filter paper, burned separately, is added and the 

compound heated gently at first, the crucible being covered until the ammonia 

is driven off, and then more strongly until the mass is snow white. The residue 

is cooled in a desiccator and weighed as Mg2P20/. The ammonium magnesium 

phosphate may be filtered directly into a weighed Gooch crucible and ignited, 

thus avoiding the carbon of the filter paper, and shortening the period of ignition, 

less heat being required to obtain the white magnesium pyrophosphate. 

» F. A. Gooch and M. Austin, Am. Jour. Sci. (4), 7, 187, 1899. W. Gibbs, C. N. 28, 
51, 1873. H. Struve, Zcit, anal. Chem., 36, 289, 1897. 

« Z. anal. Chem., 612, 19(K). =» Am. Jour. Sci. (3), 6, 114, 1873. 

* Details of the two methods maybe found inTreadwell&Hall,"Analytical Chemistry." 



256 MAGNESIUM 

Factors.! MgJPAXO.3621 =MgO or 0.2184 =Mg or X0.7572 

=MgCO,or X 1.0811 =MgS04 or X 2.2143 =MgS04-7H,0. 

Notes on Magnesium 

The ignition is conducted gently at first to gradually oxidize the carbon that the 
precipitate contains. With rauid ignition the particles are inclosed in the mass in a 
form that it is almost impossible to completely oxidize, so that the final residue is 
gray instead of white. L. L. de Koninck ' considers that the blackening of the precipi- 
tate is frequently due to the presence of organic bases in commercial ammonia and 
its salts, rather than to the fibers of filter paper occluded in the mass. With caution, 
the filter and residue may be ignited wet, the heat being low until the filter completely 
chars and then being increased, with the cover removed, until the residue is white. 

Impurities. The precipitate mav contain traces of lime that remained soluble 
in ammonium oxalate. This may be determined by dissolving the pyrophosphate 
in dilute sulphuiic acid followed by addition of 9 to 10 volumes of absolute alcohol. 
Calcium sulphate, CaS04. precipitates and settles out on standing several hours. It 
may be filtered off, dissolved in hydrochloric acid and precipitated as oxalate in the 
usual way and so determined. 

A residue remaining after treating the pyrophosphate with acid is generally SiOs. 

The presence of manganese may be detected by dissolving the magnesium pyro- 
phosphate, M^PsOt, in nitric acid and oxidizing with sodium bismuthate. (See 
method under Manganese.) 

Properties of Ammonium Magnesium Phosphate. Readily soluble in dilute 
acids. One hundred cc. of pure water at 10° C. will dissolve 0.0065 Kraxn. The 
presence of ammonia greatly decreases the solubility of the salt, e.g., 2.5% ammonia 
decreases the solubility to 0.00006 gram M^ per 100 cc. The presence of ammonium 
salts increase the solubility of the precipitate, e.g., 1 gram of ammonium chloride 
will increase the solubility to 0.0013 gram MgO.' 

VOLUMETRIC DETERMINATION OF MAGNESIUM 

Titration of the Ammonium Magnesium Phosphate with 

Standard Acid 

The procedure known as Handy's volumetric method for magnesium,* depends 
upon the reaction MgNH4P044-H,S04=MgS044-NH4H2P04. An excess of 
standard sulphuric acid is added to the precipitate and the excess of acid titrated 
back with standard ammonium hydroxide. 

Procedure. The method of precipitation of the magnesium ammonium 
phosphate is the same as has been described under the gravimetric method. The 
precipitate is washed several times by decantation with 10% ammonium hydrox- 
ide solution (1 part NH4OH, sp.gr. 0.90 to 9 parts water), and finally on the 
filter. After draining, the filter is opened out, the moisture removed as much as 
possible by means of dry filter papers. The residue may he dried in the room 
for about forty-five minutes or in the air oven at 50 to 60° C. for fifteen to 
twenty minutes.' When the filter has dried, anunonia will have been expelled. 
The substance is placed in a dry beaker, N/10 sulphuric acid added in excess 
(methyl orange indicator), the solution diluted to 100 cc. and the excess of acid 
titrated with N/10 sodium hydroxide. 

One cc. N/10 H2SO4 =0.002 gram MgO. 

^ Based on atomic weights of 1916. 

« Zeit. analy. Chem., 29, 165, 1890. 

' Mellor, " Quantitative Inorganic Analysis," J. B. Lippincott Co., Pub. 

* James Otis Handy, Jour. Am. Chem. Soc^ 22, 31. 

' Low, " Technical Methods of Ore Analysis," Wiley & Sons, Pub. 



MANGANESE 

Wilfred W. Scott 

Mn, at.wt. 54.93; sp.gr. 7.42^; m.p. 1260"^; b.p. 1900** C ^ ; oxides, MnO, 
MnsOs, (Mn304 ignition in air), MnOs, MnOa, Mn207. 

DETECTION 

In the usual course of analysis manganese is found in the filtrate from the 
hydroxides of iron, aluminum and chromium, the previous groups having been 
removed with hydrochloric acid, hydrogen sulphide and ammonium hydroxide in 
presence of ammonium chloride. Manganese, cobalt, nickel and zinc are pre- 
cipitated as sulphides in an ammoniacal solution. The sulphides of manganese 
and zinc are dissolved by cold dilute hydrochloric acid, H2S expelled by boiling 
and manganese precipitated as the hydroxide by addition of potassium hydroxide 
in sufficient amount to dissolve the zinc (sodium zincate). Manganese is now 
confirmed by dissolving this precipitate in nitric acid and adding red lead or 
lead peroxide to the strong nitric acid solution. A violet-colored solution is pro- 
duced in presence of manganese. Chlorides should be absent. 

Manganese in soils, minerals, vegetables, etc., is detected by incinerating 
the substance, treating the ash with nitric acid and taking to dryness, the residue 
is taken up with water and the mixture filtered. To the filtrate is added a few 
cc. of 40% ammonium persulphate and a little 2% silver nitrate solution. A 
pink color is produced in presence of manganese. 

Manganese compounds heated with borax in the oxidizing fiame produce 
an amethyst red color. The color is destroyed in the reducing flame. 

Fused with sodium carbonate and nitrate on a platinum foil manganese 
compounds produce a green-colored fusion (" robin egg blue")* 

ESTIMATION 

Manganese may be determined accurately gravimetrically or volumetrically. 
The former methods may be used for high-grade manganese ores, the latter are 
generally preferred for determining manganese in steel and in alloys and are 
applicable to a wide range of substances. 

The most important ore of manganese is pyrolusite, MnOj. Other ores are 
braunite, Mn203; hausmannite, MuaOi; manganite, MnjOs-HjO; albanite, 
MnS; haurite, MnSx; dialogite, MnCO,; rhodonite, MnSiOj. 

Speigeleisen or ferromanganese is an important alloy for the steel industry. 
In addition to the requirement of the element in the analysis of the above sul3- 
stances it is determined in certain paint pigments — green and violet manganous 
oxides, in dryers of oils, etc. It occurs in a number of alloys. 

^Van NoBtrand's Chem. Annual — Olsen. 

* Circular 35 (2d Ed.) U. S. Bureau of Standards. 

257 



258 MANGANESE 



Preparation and Solution of the Sample 

In dissolving the sample the following facts will be recalled: The metal 
dissolves in dilute acids, forming manganese salt«. The oxides and hydroxides 
of manganese are soluble in hot hydrochloric acid. Manganous oxide is soluble 
in nitric or in sulphuric acid; the dioxide is insoluble in dilute or concentrated 
nitric acid, but is soluble in hot concentrated sulphuric acid. 

Ores of Manganese. A sample of powdered ore weighing 1 gram is brought 
into solution by digesting with 25 to 50 cc. of strong hydrochloric acid for fifteen 
to thirty minutes on the steam bath. If much silica is present 5 to 10 cc. hydro- 
fluoric acid will assist solution. Five cc. of sulphuric acid are added and the 
mixture evaporated and heated until fumes of sulphur trioxide are evolved. 
The residue is taken up with a little water and warmed until the sulphates have 
dissolved. If decomposition is incomplete and a colored residue remains, this 
is filtered off, ignited in a platinum dish and fused with a little potassium bisul- 
phate. The fusion is dissolved in water containing a little nitric acid and the 
solution added to the bulk of the sample. 

If manganese is to be determined volumetrically the removal of iron is not 
necessary. If, however, a gravimetric procedure is to be followed, iron and 
alumina are removed by the basic acetate method given under separations 
and manganese precipitated in the filtrate. In presence of small amounts of 
iron and alumina, precipitation with ammonia in presence of ammonium chloride 
will remove these elements without appreciable loss of manganese, a double 
precipitation being usually advisable. For volumetric procedures in ores con- 
taining over 2% manganese an aliquot portion of the sample is taken for the 
determination. The portion should not contain over 0.01 gram of manganese. 

Sulphide Ores — Pyrites, etc. The sample is either roasted to oxidize the 
sulphide and then dissolved in hydrochloric acid as above stated or it is treated 
according to the procedure given for iron pyrites under sulphur. 

Slags. These may be decomposed with hydrofluoric and hydrochloric acid 
with final expulsion of these acids with sulphuric acid. Manganese is best deter- 
mined in the extract by a volumetric method. 

Iron Ores. The treatment is the same as that recommended for ores of 
manganese. The residue remaining upon evaporation with sulphuric acid is 
dissolved in a little water and about 30 cc. of nitric acid (sp.gr. 1.135) added. 
Manganese is now determined by the bismuthate method. 

Alloys. Manganese Alloys. One gram of ferromanganese is dissolved in 
50 cc. of dilute nitric acid (sp.gr. 1.135) and oxidized with sodium bismuthate 
with boiling. The cooled solution is diluted to 500 cc. and 10 to 25 cc. is treated 
with about 30 cc. of dilute nitric acid and manganese determined by the bis- 
muthate method. The amount of sample taken is governed by the manganese 
content. This should not exceed 0.01 gram of the element if the volumetric pro- 
cedure is to be followed. 

Manganese Bronze. Five grams of drillings are dissolved in dilute nitric 
acid (1.2), in a large beaker, using only sufficient acid to cause solution. If much 
free acid is present evaporation to small volume is necessary to exjx^l the nitric 
acid. The coiicentrate is dihited to 200 cc. and hydrogen sulphide passed in to 
precipitate? copijcr. The solution is diluted to 250 cc. and 50 cc. filtered off 
( = 1 gram). The HjS gjus is expelled by boiling, the solution being concentrated to 
about 15 cc. Twenty-five cc. of nitric acid are added and manganese precipitated 



MANGANESE 259 

by adding potassium chlorate in small portions. The chlorine is boiled off and the 
precipitate filtered onto asbestos and washed with concentrated nitric acid. This 
is now determined volumetrically by treating with an excess of ferrous sulphate 
of known strength and titrating the excess with standard permanganate. 

Feiro-titamum Alloy. This is best decomposed by fusion with sodium carbon- 
ate, to which a pinch of sodium peroxide has been added. The fusion is extracted 
with water and the residue containing iron, manganese and nickel filtered onto 
asbestos. Manganese is dissolved in 25 to 30 cc. of nitric acid by treating with 
SOj gas or hydrogen peroxide and manganese determined by the bismuthate 
method. 

Feiro-chromiumy Metallic Chromium. These are best decomposed by fusion 
with sodium peroxide {fi\e times the weight of sample taken), the fusion being 
made in a nickel crucible. The treatment is now the same as that recom- 
mended for ferro-titanium. 

Ferro-aluminumy Vanadium Alloys. The method used for steel is suitable 
to either of these substances. 

Molybdenum Alloys. The alloy is decomposed with hydrochloric acid, and 
iron separated by the basic acetate method, a large excess of acetate being used. 
Manganese is precipitated as the dioxide by means of bromine and ammonia by 
the detailed procedure given later. Manganese is dissolved in nitric acid after 
reduction in the acid solution by addition of a little sodium thiosulphate or SO2 
gas. It is now oxidized to permanganate by means of red lead and determined 
either colorimetrically or by titration with a standard solution of sodium arsenite. 

Tungsten Alloys. These are best decomposed by treating 1 gram of the 
substance with 5 to 10 cc. of hydrofluoric acid and a few cc. of strong nitric acid 
and digesting until the solution is complete. The hydrofluoric acid is expelled * 
by taking to dryness, a few drops of sulphuric acid having been added. The 
residue is taken up with water and boiled with SOi water. The solution is 
made to definite volume and manganese determined volumetrically on an aliquot 
portion. 

Silicon Alloys. One gram of the alloy is treated with 50 cc. of dilute nitric 
acid (sp.gr. 1.2) and 5 cc. of hydrofluoric acid. The graphite is filtered off and 
the hot solution treated with sodium bismuthate and kept boiling for about fifteen 
minutes after the manganese dioxide has been precipitated. The bismuthate 
method for estimating manganese is recommended. 

Iron and Steel. 0.5 to 1 gram of steel is dissolved by heating with 30 to 
50 cc. of dilute nitric acid (1.135). The volumetric method by oxidation with 
sodium bismuthate is generally recommended, no separations of other substances 
being required, as manganese may be determined directly in the sample. 

Pig Iron. One gram of the drillings is dissolved in 30 cc. of dilute nitric acid 
(1.135 sp.gr.), and as soon as the action has ceased the sample is filtered through a 
7-cm. filter and the residue washed with 30 cc. more of the acid. The filtrate 
containing the manganese is now treated according to the procedure for steel. 

^ Brearly and Ibbotson state that although neither tungsten nor hydrofluoric acid 
interfere with the bismuthate method of determining man^^anese, the two combined 
lead to erratic results, henoe the removal of hydrofluoric acid is necessary. 



260 MANGANESE 



SEPARATIONS 

This section includes methods of special separations of manganese from 
elements that may interfere in its determination. As is frequently the case, 
isolation of manganese is not necessary, since it may be determined volumetri- 
cally in presence of a number of elements, which would interfere in its gravimetric 
determination. The analyst should be sufficiently familiar with the material 
to avoid needless manipulations, which not only waste time, but frequently lead 
to inaccurate results. 

Removal of Elements of the Hydrogen Sulphide Group. This sepa- 
ration may be required in the analysis of certain alloys where a separation of 
manganese from copper is required. 

The acid solution containing about 4% of free hydrochloric acid (sp.gr. 1.2), 
is saturated with hydrogen sulphide and the sulphides filtered off. Manganese 
passes into the filtrate. This treatment will effect a separation of manganese 
from mercury, lead, bismuth, cadmium, copper, arsenic, antimony, tin and the 
less common elements of the group. 

Separation of Manganese from the Alkaline Earths and the Alkalies. 
The separation is occasionally required in the analysis of clays, limestone, 
dolomite, etc. It is required in the complete analysis of ores. In the usual 
course of a complete analysis of a substance, the filtrate from the hydrogen 
sulphide group is boiled free of HjS and is treated with a few cc. of nitric acid 
to oxidize the iron. The solution is made slightly ammoniacal with ammonia, 
in presence of ammonium chloride, whereby iron, aluminum and chromimn are 
precipitated as hydroxides. The filtrate is treated with hydrogen sulphide or 
colorless ammonium sulphide, whereby manganese, nickel, cobalt and zinc are 
thrown out as sulphides and the alkaline earths and alkalies remain in solution. 

Separation of Manganese from Nickel and Cobalt 

The free acid of the sulphate or chloride solution of the elements is neutralized 
with sodium carbonate and a slight excess added. It is now made strongly acid 
with acetic acid and 5 grams of ammonium acetate added for every gram of nickel 
and cobalt present. The solution is now diluted to about 200 cc. and saturated 
with hydrogen sulphide, whereby nickel and cobalt are precipitated as sulphides 
and manganese remains in solution. 



Separation of Manganese from Iron and Aluminum, Basic 

Acetate Method 

The procedure effects a separation of iron, aluminum, titanium, zirconium 
and vanadium from manganese, zinc, cobalt and nickel. 

The separation depends upwn the fact that solutions of acetates of iron, 
aluminum, titanium, zirconium and vanadium are decomposed when heated 
and the insoluble basic acetates precipitated, whereas the acetates of manganese, 
zinc, cobalt and nickel remain undecomposed when boiled for a short time. 

Fe(C2H,0,),+2HOH =2HC JI,0,+Fe(OH), • C2II3O2. 



MANGANESE 261 

The solvent action of the liberated acetic acid is prevented by the addition of 
sodium acetate * which checks ionization of the acid. The method requires care 
and is somewhat tedious, but the results attained are excellent. 

Procedure. To the cooled acid solution of the chlorides is added a concen- 
trated aqueous solution of sodium carbonate from a burette with constant stir- 
ring* until the precipitate that forms dissolves slowly. A dilute solution of the 
carbonate is now added until a slight p>ennanent opalescence is obtained. With 
the weak reagent and careful addition of the carbonate drop by drop the proper 
neutralization of the free acid is obtained. With considerable iron present the 
solution appears a dark red color, fading to colorless as the quantity of iron 
decreases to a mere trace in the solution. Three cc. of acetic acid (sp.gr. 1.044) 
are added to dissolve the slight precipitate. The more perfect the neutralization 
before heating the less amount of reagent required for precipitating iron — an 
excess of reagent does no harm. If this does not clear the solution in two 
minutes, more acetic acid is added a drop at a time until the solution clears, 
allowing a minute or so for the reaction to take place with each addition. The 
solution is diluted to about 500 cc. and heated to boiling and 6 cc. of a 30% sodium 
acetate solution added. The solution is boiled for one minute and removed 
from the iSame. (Longer boiling will form a gelatinous precipitate, difficult to 
wash and filter.) The precipitate is allowed to settle for a minute or so, then 
filtered, while the liquid is hot, through a rapid filter and washed with hot, 5% 
sodium acetate solution three times. The apex of the filter is punctured with a 
glass stirring rod and the precipitate washed into the original beaker in which 
the precipitation was made with a fine stream of hot, 1 : 1 hydrochloric acid 
solution from a wash bottle. (Dilute HNOs may be used in place of HCl.) 

A second precipitation with neutralization of the acid and addition of sodium 
acetate is made exactly as directed above. It is advisable to evaporate the solu- 
tion to small volume to expel most of the free mineral acid before addition of 
NatCOj to avoid large quantities of this reagent. The filtrates contain man- 
ganese, zinc, cobalt and nickel; the precipitate iron, aluminum, titanium, zirco- 
nium, vanadixmx. 

Separation of Manganese as the Dioxide, Mn02 

The procedure is of special value in the complete analysis of ores where a basic 
acetate separation of iron and aluminum has been made, and a gravimetric esti- 
mation of other constituents in the solution are desired. 

The procedure depends upon the principle that manganese in a dilute solu- 
tion of manganous salt is oxidized to manganese dioxide and so precipitated, 
when boiled with bromine or certain other oxidizing agents: 

MnCl,+Br2+2H20 =MnO,+2HCl+2HBr. 

The free acid formed by the reaction nmst be neutrahzed either by ammonia 
or by the presence of a salt of a weak acid such as sodium acetate, other v\'ise 
the precipitation of manganese will be incomplete. In presence of ammonium 
salts much of the bromine is used up reacting with ammonia, 

MnCl,+Br,+3NH3+2H20 = MnO,+2XH4Cl+NH4Br-fHBr. 
* Sodium acetate is preferred to ammonium acetate, though the latter may be used. 



262 MANGANESE 

At the same time an acid is fonned, which reacts with the free ammonia. It 
is necessary to have the solution ammoniacal throughout the reaction to pre- 
vent resolution of the manganese. 

Procedure. To the solution containing manganese is added 4 to 5 grams of 
sodium acetate (unless already present in excess), the solution being diluted 
to about 200 cc. Bromine water is added until a distinct color of bromine is 
evident. The mixture is boiled and kept boiling for ten to fifteen minutes, 
additional bromine being added in small portions. The precipitate is allowed 
to settle and filtered off. The filtrate is boiled with additional bromine to 
ascertain whether the manganese has been completely removed from the solution. 

If ammonia is present, as is frequently the case, it is advisable to add more 
of the reagent from time to time, the solution having a distinct odor of ammonia 
after the last portion of bromine has been added. When large amounts of 
manganese are present, several separations may be required to remove the element 
from the subsequent filtrates. 

The precipitated dioxide may be dissolved in sulphuric acid and manganese 
determined volumetrically or gravimetrically. 

It may be ignited directly and weighed as Mn304. 

It may be evaporated with sulphuric acid and manganese determined as 
MnS04. 

Manganates of zinc or calcium will be precipitated if present in large amounts. 

Manganese may also be precipitated by ammonium persulphate in an ammoniacal 
solution, potassium chlorate and chloride of lime in presence of zinc chloride in a 
neutral solution.* 



GRAVIMETRIC METHOD 
Determination of Manganese as Pyrophosphate 

Manganese is precipitated as ammonium manganese phosphate, NH4MnP04, 
and then ignited to pyrophosphate, Mn2P207. The method is known as Gibbs' 
Phosphate Process.* 

Ptocedure. The cold solution of manganese chloride ' obtained as directed 
in previous sections, should be diluted so as to contain not over 0.1 gram of 
manganese oxide equivalent per 100 cc. of solution. A cold saturated solution 
of anmionium sodium phosphate (microcosmic salt, 170 grams per liter; 9 cc. 
precipitates an equivalent of 0.1 gram of the oxide) is now added in shght excess. 
The solution is made strongly ammoniacal and heated to boiling, the boiling being 
continued until the precipitate becomes cr>\stalline. After allowing to settle 
until cold, the precipitate is filtered off (the filtrate being tested with more of the 
precipitating reagent to assure that an excess had been added), and dissolved 
in a little dilute hydrochloric or sulphuric acid. 

Reprecipitation of the phosphate. The free acid is neutralized with anmionia 
added in slight excess until the odor is quite distinct, the solution heated to 
boiling, and a few cc. of additional phosphate reagent added. The cr^^stalline 

1 J. Pattinson's Method, Jour. Chcm. Soc., 35, 3G5, 1899. 
'C.ibbs'C.X., 17, 195, 18GS. 

' Some analysts prefer to add the phosphate reagent to the strongly ammoniacal 
solution, boiling hot. 



MANGANESE 263 

precipitate is filtered into a weighed Gooch crucible, washed free of chlorides 
with very dilute ammonia (AgNOa+HNO, test), dried and ignited to the pyro- 
phosphate. The ignition is conducted, as in case of magnesium, by heating first 
over a low flame and gradually increasing the heat to the full power of the 
burner. The final residue will appear white or a pale pink. 

MnJPAX 0.4996 =MnO, 
MuiPjOtX 0.3869 =Mn. 

Notes. Zinc, nickel, copper and other elements precipitated as phosphates should 
be absent from the solution. The separation from iron is generally made by the basic 
acetate method and manganese precipitated from the filtrate, free of other elements, 
as the peroxide MnOs, by means of bromine added to the ammoniacal solution. Other 
oxidising reagents may be used, as has been stated. The dioxide is dissolved in 
strong hydrocUoric acid and the above precipitation effected. 



VOLUMETRIC METHODS 

Bismuthate Method for Determination of Manganese ^ 

The method is based on the fact that a manganous salt in the presence of an 
excess of nitric acid is oxidized to permanganic acid by bismuth tetroxide. 'The 
permanganic acid formed is very stable in nitric acid of 1.135 sp.gr., when the solu- 
tion is cold, but in hot solutions the excess of the bismuth tetroxide is rapidly 
decomposed, and then the nitric acid reacts with the permanganic acid, and, as 
soon as a small amount of manganous salt is formed, the remainder of the per- 
manganic acid is decomposed, manganous nitrate dissolves, and manganese 
dioxide precipitates. 

In the cold, however, the excess of the bismuth salt may be filtered off, and 
to the clear filtrate an excess of ferrous sulphate added, and the amount necessary 
to deoxidize the permanganic acid determined by titrating with permanganate. 
The end reactions are very sharp and the method is extremely accurate, but the 
presence of even traces of hydrochloric acid utterly vitiates the results. As 
pointed out by Reddrop and Ramage, bismuth tetroxide, which was used by 
Schneider, is difficult to obtain free from chlorides, and they recommended sodium 
bismuthate, which they prepare as follows: Heat 20 parts of caustic soda nearly 
to redness in an iron or nickel crucible, and add, in small quantities at a time, 
10 parts of basic bismuth nitrate, previously dried in a water oven. Then 
add 2 parts of sodium peroxide and pour the brownish-yellow fused mass on an 
iron plate to cool; when cold, break it up in a mortar, extract with water, and 
collect on an asbestos filter. The residue, after being washed four or five times 
by decantation, is dried in the water oven, then broken up and passed through 
a fine sieve. 

Nitric add (sp.gr. 1.135). A mixture of 3 parts of water and 1 part of 
strong nitric acid. 

Nitric acid (3%). Thirty cc. of strong acid to the liter. 

Permanganate Solution and Ferrous Sulphate Solution. One gram of 
potassium permanganate to the liter gives a solution of convenient strength, 
and 12.4 grams of ferrous ammonium sulphate and 50 cc. of strong sulphuric acid, 

*A. A. Blair, "Chemical Analysis of Iron,'* J. B. Lippincott Co., Pub. 



264 MANGANESE 

made up to 1 liter, gives a solution which is almost exactly equal to the per- 
manganate solution. As the strength of the ferrous sulphate solution changes 
quite rapidly while the permanganate remains imaltered for months, it is 
unnecessary and troublesome to keep them of the same strength. By using a 
constant volume of the ferrous sulphate solution and testing it against the per- 
manganate solution every day, the calculation of the results is very simple. 
It is necessary that the conditions should be the same in getting the strength 
of the ferrous solution as in titrating a solution for manganese, and after many 
experiments the following method was adopted: Measure into a 200-cc. flask 
50 cc. of nitric acid (1.135), cool, and add a very small amount of bismuthate, 
dilute with 60 cc. of 3% nitric acid, filter into a 300-cc. flask, and wash with 60 cc. 
of 3% nitric acid. If the felt is well coated with bismuthate it is unnecessary to 
add any to the nitric acid in the flask, as filtration through the mass of bis- 
muthate on the felt will answer the purpose. Run in from the pipette (see Fig. 
46) 25 cc. of ferrous sulphate solution and titrate with the permanganate to a faint 
pink. This gives the value in permanganate of the ferrous sulphate solution. 

The permanganate solution may be standardized in three ways: 

First, by getting its value in iron, in tha usual way, and calculating its value 
in manganese. The proportion is 5Fe : Mn or 279.2 : 54.93 =0.1967. 

Second, by titrating a steel containing a known amount of manganese and 
getting the value of the solution by dividing the percentage of manganese by the 
number of cc. of the permanganate used. 

Third, by making a solution of pure manganese sulphate and determining 
the manganese in it by evaporating a weighed amount of the solution to dryness, 
heating to dull redness, and weighing as manganese sulphate, which, multiplied 
by 0.36377, gives the amount of manganese. Five grams of C.P. manganese sul- 
phate dissolved in 500 cc. of water and filtered will give a solution containing 
about 0.0035 gram of manganese to the gram of solution. Weigh 1 to 3 grams 
of the solution in a crucible, transfer to a 200-cc. flask, using 50 cc. of nitric acid 
(sp.gr. 1.135), cool, and add 0.5 to 1 gram bismuthate, and allow it to stand for 
three or four minutes, shaking at intervals. Add 50 cc. of 3% nitric acid and 
filter through the asbestos filter and wash with 50 or 60 cc. of the same acid. 
Run 25 cc. of the ferrous sulphate solution into the flask from the pipette and 
titrate with the permanganate solution to a faint pink. Subtract the number 
of cc. of the permanganate solution obtained from the value of the 25 cc. of the 
ferrous sulphate solution in p>ermanganate, and the result is the number of cc. 
of the permanganate corresponding to the manganese in the manganese sul- 
phate solution used. Divide the weight of the manganese in the manganese 
sulphate used by the nimiber of cc. of permanganate and the result is the value 
of 1 cc. of the permanganate solution in manganese. 

Example. One gram manganese sulphate solution contains 0.003562 gram 
manganese; 2.0372 grams manganese sulphate solution equal 0.0072565 gram 
manganese; 25 cc. ferrous sulphate solution equal 24.5 cc. pennanganate solution; 
2.0372 grams manganese sulphate solution, after oxidation and addition of 25 cc. 
ferrous sulphate solution, require 3.6 cc. permanganate solution; 24.5 cc. —3.6 cc. 
=20.9 cc; 0.0072565 divided by 20.9=0.0003472, or 1 gram pennanganate 
equals 0.0003472 gram manganese. If, then, 1 gram of steel, after oxidation 
and addition of 25 cc. ferrous sulphate solution, requires 6.2 cc. permanganate 
solution to give the pink color, 24.5-6.2 = 18.3X0.0003472=0.006354 gram, 
or the sample contains 0.635% manganese. 



MANGANESE 



265 



Procedure. The nitric acid Bolution of the sample placed in a 200-cc. 
Erlenmeyer flask is treated aa follows: 

Cool, and add about 0.5 gram of biamuthate. The bLsnuithute may be 
measured in a smaU spoon, and experience will soon enable the operator to judge 
of the amount with sufficient accu- 
racy. Heat for a few minutes, or 
until the pink color has disappeared, 
with or without the precipitation of 
manganese dioxide. Add sulphurous 
acid, solution of ferrous sulphate, or 
sodium thioBulphate, in sufHcient 
amount to clear the solution, and 
beat until all nitrous oxide has been 
driven off. Cool to about 15° C, 
add an excess of bismuthate, and 
agitate for a few minutes. Add 60 cc. 
of water containing 30 cc. of nitric 
acid to the liter, and filter through 
an asbestos felt on a platinum cone 
into a 300-cc. Erlenmeyer f1a.sk, using 
suction (see Fig. 45), and wash with 
50 to 100 cc. of the same acid. Han 
into the flask from a pipette (Fig. 46) ^°- *6. Fio. 46. 

a measured volume of ferrous sulphate 

solution and titrate to a faint pink color with permanganate. The number of 
cc. of the permanganate solution obtained, subtracted from the number corre- 
sponding to the volume of ferrous sulphate used, will give the volume of per- 
manganate equivalent to the manganese in the sample, which, multiplied by the 
value of the permanganate in manganese, gives the amount of manganese in the 
sample. 

Note. In the anatysis of white irons it may be neiressary to treat the solution 
• several times with biamuthate to destroy the combined carbon. The solution, when 
cold, should be nearly colorless; if not, another treatment with biamuthate is necessary. 

Notes and Precautions 




that 

acid, which is reduced by the addition of sulnhuruus arid, the o>udatioa fiioceeda so 
slowly in cold solutions thnt if there is no delay in the hltration and titration the 
results are not affecled. Ktccia containing tungsten are sometimes troublesome on 
account of the necessity for getting rid of the tungstic acid. Those that decompose 
readily in nitric acid may be filtered and the 61trate treated like pig iron, but when 
it is necessary to use hydrochloric acid it is best to treat with aqua regia, evaporate 
to dryness, redissolve m hydrochloric acid, add a few droiis of nitric acid, dilute, 
boil, and filter. Get rid of every trace oi hydrochloric acid by repeated evajmrations 
with nitric acid, and ])roceed ns with an ordinary steel. 

The delicacy of the reaction of manganese in nitric acid solution with sodium 
bisrouthate is extraordinary; O.OCNNIOd gram of manganese gave an appreciable color 
in 50 cc, of solution. 

As will be seen in the de8iTii>(ion of the various methods of solution, the use oi 
hydrochloric acid has been avoided, because the iircsence of even traces of this 
reagent is fatal to the accurary of the method. Where it is impossible to avoid its 
use, and its presence is susjiectcd in the final nitric acid solution, the addition of a 



266 MANGANESE 

drop or two of silver nitrate will overcome the difficultyi but the filter must be 
rejected after using it for filtering a solution so treated. 

Any form of aso^tos filtering tube may he used for filtering off bismuthate, but 
the perforated cone with bell jar is the most satisfactory, because it has the largest 
area of filtering service. One filter may be used for fifty or more determinations, 
and the time occupied in filtering and washing one determination is only from one 
minute and a half to three minutes. The filtrate must be clear, for the least par- 
ticle of bismuthate carried through will vitiate the result by reacting with the excess 
of ferrous sulphate. As soon as the filtration and washing are completed, the ferrous 
sulphate should be added, and the excess titrated with the permanganate solution, 
as the permanganic acid gradually decomposes on standing, and the warmer the 
solution the more rapid is the decomposition. At a temperature of 5° C. the solu- 
tion will remain \maltered for several nours, but at 40® C, fifteen minutes will show 
an appreciable change. The larger the amount of manganese the more rapid the 
change. 

It is especially important not to allow the solution to stand after adding the 
ferrous sulphate, as the excess of this reagent reacts with the nitric acid in a few 
minutes and the formation of the smallest amount of nitrous oxide is fatal to the 
accuracy of the determination. For this reason it is important to boil off every trace 
of nitrous oxide when, in the earUer part of the operation, sulphurous acid or other 
deoxidizing agent is added. 

When working with steels of unknown manganese content, it may often happen 
that 25 cc. of ferrous sulphate solution are insufficient to entirely reduce the perman^ 
sanic acid, in which case an additional amount of ferrous sulphate must be added. 
It will be noticed that the solution of permanganic acid upon the addition of an 
insufficient amount of ferrous sulphate does not necessarily retain its pink or purple 
color, but usually changes to a dirty brown. When this occurs 10 cc. more of ferrous 
sulphate is added to the flask and tne value of the two additions taken as the amount 
from which the number of cc. of permanganate, corresponding to the excess of fer- 
rous sulphate, must be subtracted. When the sample is low in manganese, the 10 cc. 
ferrous sulphate alone may be used. 

These is no advantage in using permanganate solutions differing in strength from 
the one given above, but the strength of the ferrous sulphate solution may be 
changed to meet special cases. 

Volhard's Method for Manganese ^ 

The method is based on the principle that when potassium permanganate is 
added to a neutral manganese salt all of the manganese is oxidized and pre-, 
cipitated. When this stage is reached any excess of permanganate is imme- 
diately evident by the color produced. The calculation of results may be based 
on the reaction, 

3MnS04+2KMn04+2H,0=5MnO,+K,S04+2H2SO«, 
or 

5ZnS04+6MnS04+4KMn04+14H,0=4KHS04+7H^04+5ZnH,.2MnO,, 

the ratio in either case being 2KMn04=3Mn. 

Procedure. The material decomposed with hydrochloric and nitric acid 
and taken to fumes with sulphuric acid, as stated for the preparation of the 
sample, is cooled and boiled with 25 cc. of wat^r until the anhydrous ferric 
sulphate has dissolved and continue as follows: Transfer the mixture to a 500-cc. 
graduated flask and add an emulsion of zinc oxide in slight excess to precipitate 
the iron (C.P. ZnS04 precipitated by KOII added to slight alkalinity. The 
washed precipitate is kept in a stop|)ered bottle with suflicient water to fonn 
an emulsion). 

1 A. H. Low, "Technical Methods of Ore Analysis," 7th I^xi., John Wiley & Sons, 
Pub. (See procedure for Analysis of Spiegel Iron.) 



MANGANESE 267 

Agitate the flask to facilitate the precipitation and see that a shght excess of 
zinc oxide remains when the reaction is complete. Now dilute the contents of 
the flask up to the mark with cold water, mix thoroughly and allow to stand a 
short time and partially settle. By means of a graduated pipette draw off 
100 cc. of the clear supernatant liquid and transfer it to an 8-oz. flask. While 
the precipitate in the 500-cc. flask may appear large, it actually occupies but a 
very small space, and any error caused by it may consequently be neglected. 
Likewise the error in measurement due to change of temperature during the 
manipulation is insignificant. Heat the solution in the small flask to boiling, 
add two or three drops of nitric acid (which causes the subsequent precipitate 
to settle more quickly) and titrate with a standard solution of potassium per- 
manganate. The permanganate causes a precipitate which clouds the liciuid 
and it is therefore necessary to titrate cautiously and agitate the flask after 
each addition, and then allow the precipitate to settle sufficiently to observe 
whether or not the solution is colored pink. A little experience will enable 
one to judge by the volume of the precipitate formed, about how rapidly to 
run in the permanganate. The final pink tinge, indicating the end of the reac- 
tion, is best observed by holding the flask against a white background and 
observing the upper edges of the liquid. When this point is attained, bring the 
contents of the flask nearly to a boil once more and again observe if the pink 
tint still persists, adding more permanganate if necessary. In making this end- 
test avoid actually boiling the liquid, as a continual destruction of the color may 
sometimes thus be effected and the true end-point considerably passed. When 
the color thus remains permanent the operation is ended. Observe the number 
of cc. of permanganate solution used and calculate the result. 

It is customary to use the same permanganate solution for both iron and man- 
ganese. Having aetcrmincd the factor for iron, this may be multiplied by 0.2952 
to obtain the factor for manganese. It will be observed that 2KMn04 are required 
for 3Mn, and in the reaction for iron that 2KMn04 are required for lOFe. 1 here- 
fore 658.4 parts of iron are equivalent to 164.79 parts of manganese, or, 1 part of iron 
to 0.2951 part of manganese. 

3MnS04 -f 2KMn04 +2H2O - 5Mn02 -I-K2SO4 +2H,S04, 

10FeSO4+2KMnO4 +8H2SO4 = 5Fe2(S04)8 -|-2MnS04 -f K2SO4 +8H2O. 

Ammoniuin Persulphate Method for Determining Small Amounts 
of Manganese by Colorimetric Comparison or by Titration 

The process depends upon the oxidation of manganous salts to perman- 
ganate by ammonium persulphate in presence of a catalytic agent such as silver 
nitrate: 

2Mn(NO,),+5(NH4)2S208+8HjO=5(XH4)2S04+5H,S04+4HNO,+2HMn04. 

The reaction takes place equally well in sulphuric or in nitric acid solution, 
or in a mixture of the two. The essential point is the presence of a sufficient 
amount of silver nitrate catalyst. 

Procedure. One gram of ore is dissolved in hydrochloric acid, followed by 
sulphuric and taken to fumes as directed under Preparation and Solution of the 
Sample. The sulphate taken up with water is made to a volume of 100 cc. If 
the color comparison is to be made the solution should be filtered through a 
dry filter, otherwise tlie filtration may be omitted. Twenty cc. (equal to 0.2 



268 MANGANESE 

gram) of the material is taken for the test. In the case of steel, 0.1 to 0.2 gram 
of the drilling, dissolved in dilute nitric acid, is taken. 

Oxidation. The solution is transferred to a test-tube, 1X10 ins., if the 
color comparison is to be made, or into a 150-cc. Erlenmeyer flask, if the sample 
is to be titrated. Fifteen cc. of silver nitrate solution (1.5 grams AgNOj per 
liter of water), are added; the solution heated to 80 to 90° C. by placing the 
receptacle in hot water, and then 1 gram of ammonium persulphate added. When 
the color commences to develop the sample is cooled in cold water, while the 
evolution of oxygen continues. The sample is poured into the comparison tube 
and the color compared with that obtained from an ore or steel sample of known 
manganese content, run in the same way. 

In the titration method the solution in the Erlenmeyer flask is diluted to 
about 100 cc, 10 cc. of 0.2% salt solution added, and the sample titrated with 
standard sodium arsenite until the permanganate color is destroyed. The cc. 
of the reagent used multipHed by the factor per cc. in terms of manganese 
equals weiglit of manganese in the sample titrated. 

Note. Arsenious acid reagent is made by dissolving 10 grams of arscnious oxide 
in water containing 30 grams of sodium carbonate. The solution is diluted to 1 liter. 
125 cc. of this solution are diluted to 2000 cc. This latter reagent is standardized against 
an ore or sample of steel of known manganese content, following the directions given 
under the procedure outlined. 

Oxidation of Manganese to Permanganate by Red Lead 

Red lead oxidizes manganese in nitric acid solution to permanganate. The 
method is suitable for determining this element in steel and iron in presence 
of molybdenum, aluminum, tungsten, copper and nickel, in amounts such as 
are apt to be present. The method is given in the chapter on Iron in the 
Analysis of Iron and Steel, page 227. 

ANALYSIS OF SPIEGEL IRON FOR MANGANESE ^ 

Procedure. Weigh 0.5 gram of the sample in a 250-cc. beaker, add 40 cc. 
dilute HNOj (1-2), cover with a watch-crystal, heat over Bunsen burner and 
finally expel nitrous fumes by boiling down to a small volume (5 cc.).* Wash 
into a 500-cc. graduated flask, fill about half full, neutralize with an emulsion of 
zinc oxide, adding enough to precipitate the iron and a sl'ght excess.* Dilute to 
the mark, shake well, pour into a 600-cc. beaker and mix by pouring back into the 
flisk and then into beaker. Allow the precipitate to settle, decant off two 100-cc. 
portions of clear solution into 350-cc. casseroles. Add 100 cc. water, heat to 
boiling and titrate with standard KMnO^, stirring thoroughly with heavy glass 
rod. Run in about 1 cc. at a time until the end-ix)int is passed. < Titrate the 
second pcjrtion, running it up to within 1 cc. of the end-point, and finishing a drop 
or two at a time, stirring thoroughly between each addition.* The burette 
reading gives percentage of Mn directly. 

* Procedure communicated to the author by Dr. Brcyer. 

•It is necessary to boil off nitrous fumes, as they will consume KMn()4. 

* Always test the zinc oxide for reducing substances. 

*In samples in which the T)erc<»ntage of Mn is known approximately, almost 
the full amount of KMn()4 can ne added at once. 

*Do not mistake the reflection of i)recipitated MnOj for excess of KMnOi. 
If pro])erly carried out the MnOj should collect in center of casserole. 



MANGANESE 269 

Preparation and Standardization of Permanganate.^ Dissolve 23.23 
grams C.P. KMn04 in 12 liters of distilled water, shake thoroughly and allow to 
stand a week or two before using. 

Standardization. Weigh .15 gram C.P. sodium oxalate (Bureau of Standards) 
into a 400-cc. beaker. Dissolve in 200 to 250 cc. hot water (80 to 90®), add 10 cc. 
(1:1) sulphuric acid. Titrate at once with KMnO*, until 1 drop gives a per- 
manent pink. 

When .15 gram sodium oxalate is taken, it should consume 36.87 cc. KMnO^, 
if the permanganate is of correct strength, i.e., 1 cc. =1% in .1 gram sample 
titrated. 

*The Standardization of KMn04 solution by Sodium Oxalate, McBride, J.A.C.S., 
84, 393. Miller, ** Quantitative Analysis for Mming Engineers." 



MERCURY 

Wilfred W. Scott 

Hg, at.wt. ZO0.e; sp.gr. 13.595; ^ m.p. 38.90'';' 6.p. 357.33"* C; ^ oxides, 

HgaO, HgO. 

DETECTION 

Metallic mercury is recognized by its physical properties. It is the only 
metal which is a liquid at ordinary temperatures. The element forms a convex 
surface when placed on glass. 

Mercury in the mercurous form is precipitated by hydrochloric acid as white 
mercurous chloride, HgCl. This compound is changed by ammonium hydroxide 
to the black precipitate of metallic mercury and nitrogen dihydrogen mercuric 
chloride.' 

Mercury in the mercuric form is not precipitated by hydrochloric acid. The 
sulphide of the element is thrown out from an acid solution as black HgS. 
The precipitate first appears white, changing to orange-yellow, then brown 
and finally to black, as the HjS gas is passed into the solution. The element 
is distinguished from the other members of the group by the insolubility of 
its sulphide in yellow anunonium sulphide and in dilute nitric acid. 

If the mercury sulphide is dissolved in aqua regia, the nitric acid expelled 
by taking to dryness, then adding hydrochloric acid and evaporating again to 
dryness, the residue taken up with a little hydrochloric acid, diluted with water, 
and treated with a solution of stannous chloride, a white precipitate of mer- 
curous chloride is first formed, which is further reduced to metallic mercury by 
an excess of the reagent. 

ESTIMATION 

The metal is found free in the upper portions of cinnabar deposits. As 
an amalgam with silver it occurs in horn silver. Cinnabar, HgS, is the only ore 
of mercury of commercial importance. The element has been found in quartz, 
sandstone, schists, iron pyrites, bituminous substances, eruptive and sedimentary 
rocks of all ages. It occurs in sulphide ores of other metals — especially in zinc 
ores. 

Preparation and Solution of the Sample 

It will be recalled that nitric acid is the best solvent for the metal and its 
amalgams. The oxides are insoluble in alkalies. Mercuric oxide is dissolved 
by acids. Hydrochloric acid fonns mercurous chloride with the lower oxide, 
insoluble in dilute hydrochloric acid. 

^Van Nost rand' sC hem. Annual — Olsen. 

'Circular 35 (2d Ed.) U. S. Bureau of Standards. 

• Prescott and Johnson, "(Qualitative Chemical Analysis." 

270 



MERCURY 271 

Ores. If mercury is to be determined by the dry procedure, the finely 
ground sample may be mixed directly with the flux and determined as directed 
later. 

For the wet methods the ore is decomposed in a covered porcelain dish with 
aqua regia, heating gently. The solution is evaporated to drj-ness on the water 
bath. The residue is taken up with hydrochloric acid and again evaporated to 
dryness to expel the nitric acid. The residue is again dissolved by adding a little 
hydrochloric acid. Mercury will now be in solution and may be determined by 
precipitation as mercuric sulphide by the gravimetric procedure. 

For opening up the ore for the volumetric method by Seamon see method 
at close of the chapter, page 274. 

SEPARATIONS 

Separation of Mercury from the Iron and Zinc Groups, or from the 
Alkaline Earths and the Alkalies. Mercury is precipitated as a sulphide 
from an acid solution of the mercuric salt by hydrogen sulphide, together with the 
members of the hydrogen sulphide group. Suflicient acid should be present 
to prevent the precipitation of zinc sulphide. Iron, aluminum, chromium, 
manganese, cobalt, nickel, zinc, the alkaline earths and the alkaUes remain in 
solution. 

Separation of Mercury from Arsenic, Antimony, and Tin. The sul- 
phides obtained by passing hydrogen sulphide into an acid solution, preferably 
of the chlorides, are digested with yellow ammonium sulphide solution. Arsenic, 
antimony and tin dissolve, whereas mercury sulphide remains insoluble. Sul- 
phides of the fixed alkalies dissolve mercury as well as arsenic, antimony and tin, 
so cannot be used in effecting a separation. 

Separation from Lead, Bismuth, Copper and Cadmium. These elements 
remain with mercury upon removal of arsenic, antimony and tin as their sul- 
phides are insoluble in ammonium sulphide. (CuS slightly soluble.) The pre- 
cipitated sulphides are transferred to a porcelain dish and boiled with dilute 
nitric acid, sp.gr. 1.2 to 1.3. After diluting shghtly with water the solution is 
filtered and the residue of mercuric sulpliide washed with dilute nitric acid and 
finally with wat«r. If much lead is present in the solution it is apt to contami- 
nate the residue by a portion being oxidized to lead sulphate and remaining insol- 
uble. In this case the residue is treated with aqua regia, the solution diluted and 
mercury chloride filtered from PbSO* and free sulphur. Mercury is best deter- 
mined as HgS by the anmionium sulphide method described later. Traces of 
lead do not interfere, as lead is completely removed by remaining insoluble in 
potassium hydroxide, whereas mercury sulphide dissolves. See method. 

Separation from Selenitmi and Tellurium. The mercury selenide or telluride 
is dissolved in aqua regia, chlorine water added and the solution diluted to 
600 to 800 cc, phosphorous acid is added and the solution allowed to stand for 
some time; mercurous chloride is precipitated, selenium and tellurium remain- 
ing in solution. Selenium and tellurium will precipitate in hot concentrated 
solutions when treated with phosphorous acid, but not in dilute hydrochloric 
acid solutions. 

Mercury in Organic Substances. The material is decomposed by heating 
in a closed tube with concentrated nitric acid. Mercury is now precipitated as 
a sulphide with anunonium sulphide as directed in the procedure given later. 



272 



MERCURY 



QRAVIMETRIC METHODS 

Determination of Mercury by Precipitation with Ammo- 
nium Sulphide^ 

The following method, suggested by Volhard, is generally applicable for 
determination of mercury. The element is precipitated by ammonium sul- 
phide as HgS. The precipitate dissolved in caustic is again thrown out by addi- 
tion of anmionium nitrate to the sulpho salt solution of mercury. 

Hg(SNa)t+2NH4NO,=2NaNO,+(NH4),S+HgS. 

Procedure. The acid solution of the mercuric salt is nearly neutralized by 
sodium carbonate, and is then heated with a slight excess of ammonium sulphide 

reagent, freshly prepared. Sodium hydroxide solu- 
tion is added until the dark-colored liquid begins to 
lighten. The solution is now heated to boiling and 
more sodium hydroxide added until the liquid is 
clear. If lead is present it will remain undissolved 
and should be filtered off. Anunonium nitrate is 
now added to the solution in excess and the mix- 
ture boiled until the greater part of the ammonia 
has been expelled. The clear liquid is decanted 
from the precipitate through a weighed Gooch 
crucible and the precipitate washed by decantation 
with hot water and finally transferred to the cruci- 
ble and washed two or three times more. The 
mercuric sulphide is dried at 110® C. and weighed as 
HgS. 

HgSX0.8622=Hg or X0.9o;07=HgO. 



-Z-Il-^l-V^ I Wafer] 




• €tooch ^ 
■hHgSi-S 



C5t 



Hot Wafer 



I 



Fig. 47. 



Notes. Alumina and silica are apt to be present in 
caustic. 

Free sul| hur may be removed, if present, by boiling with sodium 8ul])hite, 
Na2S(^4S = Na»S202. The sulphur may be extracted with carbon disulphide. 
The Gooch crucible is placed upon a glass tripod in a beaker, containing carbon di- 
8i;lphide, and a round-bottomed flask filled with cold water is placed over the mouth 
of the I eal er to serve as a condenser. Fig. 47. By gently heating over a water bath 
for an hour the sulphur is completely extracted from the sulphide. Carbon disul- 
phide is removed from the precipitate by washing once with alcohol followed by 
ether. The residue is now dried and weighed. 

Determination of Mercury by Electrolysis 

Mercury is readily deposited as a metal from slightly acid solutions of its salts. 

Procedure. The neutral or slightly acid solution of mercuric or mercurous 
salt is diluted in a beaker to 150 cc. with water and 2 to 3 cc. of nitric acid added. 
The solution is electrolyzed with a current of 0.5 to 0.1 ampere, and an E.M.F. 
of 3.5 to 5 volts. A gauze cathode is recommended, or a platinum dish with 
dulled inner surface may Ije used. One gram of mercury may ho de|3(jsited in 
about fifteen hours (or overnight). The time may be shortened to about three 
hours by increasing the current to 0.6 to 1 ampere. 

» Treadwell and Hall, "Analytical Chemistry," Vol. 2, 4th Ed. J. Wiley & Sons. 



MERCURY 



273 



The metal is washed with water without intemipting the current and then 
with alcohol. After remo\'ing the adhering alcohol with a filter paper, the 
cathode is placed in a desiccator containing fused potash and a snmll dish of 
mercury. The object of this mercury is to prevent loss of the deposit by 
vaporization. 

The increased weight of the cathode is due to metallic mercur>'. 

NoTEP. In the electrolysis of mercuric chloride turbidity may be caused by 
formation of mercurous chloride by reduction, but this does no harm, as the reduction 
to metallic mercury follows. 

Mercurj' may be clectrolyzed from its sulpho solutions, obtained by dissolving 
its sulphide in concentrated sodimn sulphide. 

Determination of Mercury by the Holloway-Eschka Process 

Moidified 

When mercury sulphide is heated with iron filings metallic mercury is vol- 
atilized, iron sulphide being formed. The mercurj' vapor is condensed on a 
silver or gold plate. The use of iron for this reduction was suggested by Eschka 
and his method modified by Holloway. In ores containing arsenic the addition 
of zinc oxide is recommended. Erdmann and Marchand use lime for decomposing 
the mercury compound. The reactions may be represented as follows: 

HgS+Fe=FeS+Hg or HgX+CaO=CaX+Hg+0. 

Apparatus. This consists of a deep glazed porcelain crucible, the size 
depending upon the charge of the sample to be taken. Generally a 30-cc. cru- 
cible is used for a 2-gram sample with 4 grams of flux. The crucible is covered 
by a silver or gold plate that lies 

Condenser 
CofdffMer 



perfectly flat and fits snugly 
around the edges of the crucible. 
It may be necessary to grind the 
top of the receptacle on emery 
paper to obtain a perfectly level 
edge. 

The crucible is suspended in a x 
hole through an asbestos board or 
quartz plate, to prevent the flame 
heating the upper portion of the 
vessel. 




di 



'5i her Cover 



Asbesfos Boa re/ 



i*ri9i^^ 





Crucib/e wiffi 
5amp/e 



Fig. 48. 



The lid of the crucible is kept cool by a cylindrical condenser of metal through 
which a stream of water pasvses. A small Erlenmeyer flask may be used, with 
a tube passing to the bottom of the flask through a rubber stopper, and a 
second tube just passing through the stopper. 

Holloway has a weight placed on the metal condenser to hold the lid firmly 
against the crucible. Tlie illustration (Fig. 48) shows the form of the apparatus 
set up for the run. 

Procedure. The sample containing not over 0.1 gram of mercury is placed 
in the crucible with 5 to 10 grams of fine iron filings and intimately mixed. Addi- 
tional filings are put over the charge. Sulphide ores containing arsenic are 
best mixed with about twice the weight of a flux of zmc oxide and sodium car- 
bonate in the proportion 4 to 1, and about five times the weight of iron filings 
added. 



274 MERCURY 

The weighed silver cover is placed on the crucible and the apparatus set up as 
shown in the illustration, Fig. 48. 

The bottom of the crucible is gradually heated with a small M^ker iSame 
imtil it glows slightly. Overheating should be avoided. The upper portion of 
the crucible should never become hot and the lid should remain cold. After 
heating for about thirty minutes the system is allowed to cool without discon- 
necting the condenser. The disk is now removed, dipped in alcohol and dried 
in a desiccator over fused potash or soda. The increase of weight of the dried 
disk is due to metallic mercury. 

Notes. If the sample contains less than 1 % mercury, take 2 grams; if 1 to 2% 
mercury, take 1 gram ; ii the sample contains 2 to 5%, take 0.5-gram sample. If high in 
mercury, grind sample with sand and take an aliquot portion. 

It is tidvisablc to repeat the test with a clean foil to be sure that all the mer- 
cury has been driven out of the sample. The foil may be freed from mercury by 
heating. 

VOLUMETRIC DETERMINATION OF MERCURY 
Seamon's Volumetric Method ^ 

Seamon's Volumetric Method.' "Weigh 0.5 gram of the finely ground ore 
into an Erlenineyer flask of 125 cc. capacity. Add 5 cc. of strong hydrochloric 
acid and allow it to act for about ten minutes at a temperature of about 40° C, 
then add 3 cc. of strong nitric acid and allow the action to continue for about 
ten minutes longer. The mercury should now all be in solution. Now if lead be 
present, add 5 cc. of strong sulphuric acid; it may be omitted otherwise. Dilute 
with 15 cc. of water and then add ammonia cautiously until the liquid is slightly 
alkaline. Bismuth, if present, will be precipitated. Acidify faintly with nitric 
acid, filter, receiving the filtrate in a beaker, and wash thoroughly. 

Add to the filtrate 1 cc. of strong nitric acid that has been made brownish 
in color by expasure to the light, and titrate with a standard solution of potassium 
iodide until a drop of the liquid brought into contact with a drop of starch 
liquor, on a spot-plate, shows a faint bluish tinge. It is a good plan to set aside 
about one-third of the mercury solution and add it in portions until the end- 
point is successively passed, finally rinsing in the last portion and titrating to 
the end-point very carefully. 

Deduct 0.5 cc. from the burette reading and multiply the remaining cc. used 
by the percentage value of 1 cc. in mercury to obtain the percentage in the ore. 

The standard potassium iodide solution should contain 8.3 grams of the 
salt per liter. Standardize against pure mercuric chloride. Dissolve a weighed 
amount of the salt in water, add 2 cc. of the discolored nitric acid and titrate 
as above. One cc. of standard solution will be found e(iuivalent to about 0.005 
gram of mercury, or about 1% on the basis of 0.5 gram of ore taken for assay. 

The precipitate of red mercuric iodide which fonns during the titration may 
not appear if the amount of mercury present is very small, but this failure to 
precipitate does not appear to affect the result. 

Iron, copper, bisnmth, antimony, and arsenic, when added separately to 
the ore, did not influence the results in Seamon's tests. Silver interferes. Dupli- 
cate results should check within 0.1 to 0.2 of 19c- 

^ A. H. Low. ** Technical Methods of Ore Analysis." 
* " Manual tor Assayers and Chemists," p. 112. 



MOLYBDENUM 

Wilfred W. Scott 
MOf at. wt, 96.0; sp.^r. 8.6 —9.01 ; m.p. S500'' C ; oxides, MoiOs, Mo02, MoO^ 

DETECTION 

Molybdenum appears in the hydrogen sulphide group, being precipitated by 
HjS in acid solution as the sulphide. It passes into solution by digestion with 
ammonium sulphide or sodium sulphide along with arsenic, antimony, tin, gold 
and platinum. By addition of metallic zinc, antimony, together with tin, gold 
and platinimi are precipitated as metals while molybdenum remains in solution. 
Arsenic, that has not volatilized as arsine, is expelled by evaporation. Nitric 
acid is now added and the solution taken to dryness. Molybdenum is extracted 
from the residue with ammonium hydroxide. 

A dilute solution of ammonium molybdate treated with a soluble sulphide 
gives a blue solution. 

Sodium thiosulphate added to a slightly acid solution of ammonium molyb- 
date produces a blue precipitate with a supernatant blue solution. With more 
acid a brown precipitate is formed. 

Sulphur dioxide produces a bluish-green precipitate if sufficient molybdenum 
is present, or a colored solution with small amounts. The reducing agents, stan- 
nous chloride, or zinc in acid solution, produce a play of colors when they react 
with molybdenum solutions, due to the formation of the lower oxides. The solu- 
tion becomes blue, changing to green, brown and yellow. 

Molybdenum present as molybdate is precipitated by disodium phosphate 
as yellow ammonium phosphomolybdate from a nitric acid solution. The pre- 
cipitate is soluble in ammonium hydroxide. 

A pinch of powdered mineral on a porcelain lid, moistened with a few drops 
of strong sidphuric acidy stirred and heated to fumes, then cooled, will produce 
a blue color when breathed upon. The color disappears on heating, but reappears 
on cooling. Water destroys the color. 

Molybdenite is very similar to graphite in appearance. It is distinguished 
from it by the fact that nitric acid reacts with molybdenite, MoSj, leaving a 
white residue, but has no action upon graphite. The blowpipe gives SO2 with 
molybdenite and CO2 with graphite. 

ESTIMATION 

The determination is required in the ores — molybdenite, MoSx, (60% Mo); 
molybdite, MoO» (straw yellow); wulfenite, PbMoO* (yellow, bright red, olive 
green or colorless); Ilsemannite, M0O3+M0O2; powellite, CaMo04; pateraite, 
Ck)MoO«; belonesite, MgMoOi; eosite, lead-vanado-molybdate; achromatite, 

275; 



276 MOLYBDENUM 

lead molybdate and arsenate with tin oxide and lead chloride. Some iron and 
copper ores also contain molybdenum. 

The metal is determined in certain self-hardening steels and alloys. 

The reagents ammoniiun molybdate and the oxide-molybdic acid, MoOs, 
are valuable for analytical purposes. Tests of their purity may be required. 

Preparation and Solution of the Sample 

In dissolving the substance the following facts should be kept in mind: 
The metal is easily soluble in aqua regia; soluble in hot concentrated sulphuric 
acid, soluble in dilute nitric acid, oxidized by excess to MoOs. It is dissolved 
by fusion with sodium carbonate and potassium nitrate mixture. It is insol- 
uble in hydrochloric, hydrofluoric and dUute sulphuric acids. 

The oxide MoOa is but slightly soluble in acids and alkalies; MoOs is insol- 
uble in hydrochloric and hydrofluoric acids. MoOj, as ordinarily precipitated, 
is soluble in inorganic acids and in alkalies. The oxide sublimed is difficultly 
soluble. 

Molybdates of the heavy metals are insoluble in water, the alkali molybdates 
are soluble. 

Ores. Molybdenum ores are best decomposed by fusion with a mixture of 
sodium carbonate and potassium nitrate. The cooled fusion is then extracted 
with hydrochloric acid and molybdenum determined according to one of the 
procedures described later. 

Steel and Iron. One to 2 grams of the drillings are dissolved in a mixture 
of hydrochloric and nitric acid (25 cc. HCl+1 cc. HNOj), with gentle heating. 
Additional nitric acid is added if required or potassium chlorate may be used 
to oxidize the iron. 

SEPARATION OF MOLYBDENUM FROM OTHER ELEMENTS 

Separation from Iron. Procedure in Presence of Large Amounts of 
Iron. The occurrence of molybdenum with iron and its commercial importance 
in iron materials calls for this procedure as one commonly re(iuired in the deter- 
mination of molybdenum. 

The solution is nearly neutralized with a 2N. NaOH solution, added from a 
burette cautiously, avoiding an amount that would produce a color with iron or 
form a basic molybdate. If tungsten is present or if molybdic acid has precip- 
itated in the solution or is suspected, the sample should be filtered and the 
residue treated as directed below. Sufficient 2X. XaOH to precipitate all the 
iron present in the sample (27 cc. of 2X. NaOH will precipitate 1 gram Fe) with 
about 40 cc. in excess is poured into a 500-cc. flask. If filtration is necessary, 
the paper and residue are dropped in the flask, the filter broken up, and the 
caustic heated to boiling to dissolve the molybdic acid. The solution contain- 
ing the molybdenum is also heated to boihng and added to the hot NaOH solu- 
tion, through a funnel with a constricted stem, agitating the sodium hydroxide 
during the addition. Iron hydroxide, Fe(0H)3, is precipitated free from molyb- 
denum, which remains in solution. The volume Ls made up to exactly 500 cc. 
and the precipitate allowed to settle; 250 cc. are filtered off and taken for the 
precipitation of molybdenum. Methyl orange is added as an indicator and 
the caustic neutralized with HCl. If barium, strontium, uranium, arsenic, cad- 
mium and aluminum are present, 10 to 15 cc. strong hydrochloric acid are added 



MOLYBDENUM 277 

in excess, followed by sufficient ammonium acetate to combine with the free 
mineral acid. 

This method, followed by the lead molybdate precipitation as given in the 
gravimetric methods, will effect a separation of molybdenum from barium, 
calcium, strontium, arsenic, cadmium, phosphorus, aluminum, uranium, man- 
ganese, cobalt, nickel, zinc, chromium, magnesium, mercury, copperas well as iron. 

Separation from the Alkalies. Molybdenum, precipitated as mercurous 
molybdate, by adding mercurous nitrate to the slightly acetic acid solution, or as 
molybdenum sulphide by H2S passed into the sulphuric acid solution, is separated 
from the alkalies. 

If hydrogen sulphide is passed into the sulphuric acid solution, separation of 
molybdenum from the members of the ammonium sulphide group is effected, as 
well as from members of subsequent groups. 

Separation from the Alkaline Earths. Fusion of the substance with sodium 
carbonate and extraction of the melt with water gives a solution of molybdenum, 
whereas the carbonates of barium, calcium and strontium remain undissolved 
as carbonates. 

Separation from Lead, Copper, Cadmium and Bismuth. The sulphides 
of the elements are treated with sodium hydroxide and sodium sulphide solution 
and are digested by gently heating in a pressure flask. Molybdenum dissolves, 
whereas lead, copper, cadmium and bismuth remain insoluble. If the solution 
of the above elements is taken, made strongly alkaline, and treated with HiS, 
the sulphides of the latter elements are precipitated and molybdenum remains 
in solution. The precipitates are filtered off and the filtrate containing molyb- 
denum is placed in the pressure flask, the solution made slightly acid with sul- 
phuric acid and the mixture heated under pressure, until the liquid appears colorless, 
M0S3 is precipitated and may be converted into the oxide as described later. 

Separation from vanadium is effected by a molybdenum sulphide precipi- 
tation in acid solution. 

Separation from Arsenic. Arsenic, present in the higher state of oxida- 
tion, is precipitated by magnesia mixture, added to a slightly acid solution (5 cc. 
of concentrated hydrochloric acid per 100 cc. of solution for each 0.1 gram 
arsenic). The solution is neutralized with ammonia (methyl orange), and the 
arsenic salt filtered off. M0S2 is now precipitated with H2S in presence of free 
sulphuric acid in the pressure flask. 

Separation from Phosphoric Acid. Phosphoric acid is precipitated from an 
ammoniacal solution as magnesium ammonium phosphate. Molybdenum may 
then be precipitated as the sulphide from the filtrate. 

Separation from Titanium. The metals of the ammonium sulphide group 
are precipitated by adding ammonium hydroxide and ammonium sulphide. 
Molybdenum remains in solution and passes into the filtrate. H2S is passed into 
the solution until it appears red; sulphuric acid is then added until the solution 
is acid, when molybdenum sulphide precipitates. 

Separation from Timgsten. Molybdenum, precipitated with tungsten by 
the lead molybdate method, is ignited and the mixture then treated with hydro- 
chloric acid and a few drops of nitric acid and evaporated nearly to dryness. 
Dilute hydrochloric acid is added and the solution filtered. Tungsten remains 
undissolved. 

Molybdenum may be precipitated by H2S as M0S2 in presence of tartaric 
acid. Tungsten does not precipitate. 



278 MOLYBDENUM 

GRAVIMETRIC METHODS FOR THE DETERMINATION 

OF MOLYBDENUM 

Precipitation as Lead Molybdate 

Preliminary Remarks. This method, suggested by Chatard, has been 
pronounced by Brearly and Ibbotson to be " one of the most stable processes 
found in analytical chemistry." " It is not interfered with by the presence of 
large amounts of acetic acid, lead acetate, or alkali salts (except sulphates). 
The paper need not be ignited separately and prolonged ignition at a much higher 
temperature than is necessary to destroy the paper does no harm. From faintly 
acid solution lead molybdate may be precipitated free from impurities in the 
presence of copper, cobalt, nickel, manganese, zinc, magnesium and mercury 
salts.'' It may be readily separated from iron and chromium. Barium, stron- 
tium, uranium, arsenic, cadmium and aluminum do not interfere if an excess of 
hydrochloric acid has been added to the solution followed by lead acetate and 
sufficient ammonium acetate to destroy the free mineral acid. 

The precipitate is granular, easily riltered and washed. 

Vanadium and tungsten, if present, must be removed, by separating from 
molybdenum by one of the procedures given. 

Special Reagents. Lead Acetate. A 4% solution is made by dissolving 
20 grams of the salt in 500 cc. of warm water. A few cc. of acetic acid are added 
to clear the solution. 

Precipitation of Lead Molybdate. An excess of lead acetate is added to 
the acetic acid solution, contahiing molybdenum (see Separation in Presence 
of Large Amounts of Iron), (10 cc. of 4% solution of the crystallized lead acetate 
salt will precipitate 0.01 gram of molybdenum). The solution is heated to boil- 
ing, the crystalline precipitate allowed to settle for a few minutes on the steam 
bath, then filtered hot onto an ashless filter (S. & S. No. 590 quality) and washed 
free of chlorides with hot water. 

The precipitate dried and ignited in a porcelain crucible at red heat for about 
twenty minutes is weighed as PbMoO*. 

PbMoO4X0.2615 =Mo. PbMoO4X0.3923 =MoOa. 

MoX3.8241 =PbMo04. MoO,X2.5491 ^PbMoO*. 

Determination of Molybdenum as the Oxide, M0O3 
Determination by Precipitating with Mercurous Nitrate 

Especially applicable where fusion with an alkali carbonate has been required. 

Decomposition of Ore. One gram of the ore is fused with 4 grams of fusion 
mixture, (NajCOj+KjCOj+KNOa), and the cooled melt extracted with hot water. 

If manganese is present, indicated by a colored solution, it may be removed 
by reduction with alc(jhol, the manganese precipitate filtered off and washed with 
hot water, the solution evaporated to near dryness and taken up with water, 
upon addition of nitric acid as stated below. 

The solution containing the alkaline molybdate is nearly neutralized by 
adding HNO3, the amount necessary being determined by a blank, and to the 
cold, slightly alkaline solution, a faintly acid solution of mercurous nitrate is 



MOLYBDENUM 279 

added until no further precipitation occurs. The precipitate consists of mer- 
curous molybdate and carbonate (chromium, vanadium, tungsten, arsenic and 
phosphorus will also be precipitated if present). The solution containing the 
precipitate is boiled and allowed to stand ten to fJteen minutes to settle, the black 
precipitate is filtered off and washed with a dilute solution of mercurous nitrate. 
The precipitate is dried, and as much as [X)88ible transferred to a watch-glass. 
The residue on the filter is dissolved with hot dilute nitric acid, and the solution 
received in a large weighed porcelain crucible. The solution is evaporated to 
dryness on the water bath and the main portion of the precipitate added to this 
residue, and the product heated cautiously over a low flame * until the mercury 
has completely volatilized. The cooled residue is weighed as MoOj. 

MoO,X 0.6667= Mo. 

Note. If Cr, V, W, As or P are present a separation must be effected. Molyb- 
denum should be precipitated in an H2SO4 solution in a pressure flask as the sulphide 
by HjS as given in the following method, and arsenic if present removed by magnesia 
mixture as indicated in the procedure for separation of arsenic from molybdenum. 
If these impurities are present the molybdenum oxide may be fused with a very little 
Na<COt, and leached with hot water and the filtrate treated with H2S as directed. 



Precipitation of Molybdenum as the Sulphide by H2S 

A. Precipitation from Acid Solution. By this procedure molybdenum 
18 precipitated along with members of the hydrogen sulphide group, if present, 
but free from elements of the following groups. 

The cold molybdenum solution slightly acid with sulphuric acid (in presence 
of Ba, Sr or Ca an HCl solution is necessary) is placed in a small pressure 
flask and saturated with HiS, the flask closed and heated on the water bath until 
the precipitate has settled. The solution is cooled and filtered through a weighed 
Gooch crucible. 

B. Precipitation from an Ammoniacal Solution. By this procedure moly b • 
denum is precipitated with antimony, arsenic, tin if present, but is free from 
mercury, lead, bismuth, copper and cadmium. 

Hydrogen sulphide is passed into the cold ammoniacal solution of molybdenum 
until it assumes a bright red color, it is now acidified with dilute sulphuric acid, 
the precipitate allowed to settle and the solution filtered through a weighed 
Gooch crucible. 

In either case A or B the precipitate is washed into the Gooch crucible 
with very dilute sulphuric acid followed by several washings with the acid and 
then with alcohol until free from acid. The Gooch is placed within a larger 
nickel crucible and covered with a porcelain lid. After drying at 100° C. it is 
placed over a small flame and carefully heated until the odor of SO2 can no longer 
be detected. The cover is now removed and the open crucible heated to constant 
weight. The residue consists of MoOa. 

MoO,X 0.6667= Mo. 

Note. Arsenic will contaminate the residue if present. The method for its 
romoval has been given. 

^ The oxide, MoOs, sublimes at bright red heat. 



280 MOLYBDENUM 

VOLUMETRIC METHODS FOR THE DETERMINATION 
OF MOLYBDENUM OR MOLYBDIC ACID 

The lodometric Reduction Method ^ 

Principle. When a mixture of molybdic acid and potassium iodide in pres- 
ence of hydrochloric acid is boiled, the volume having defined limits, free iodine 
is liberated and expelled and the molybdic acid reduced to a definite lower oxide; 
by titrating with a standard oxidizing agent the molybdic acid is determined. 

Reaction. 2Mo03+4KI+4HCl =2MoOJ+l2+4KCl+2H20. 

Reagents. N/10 solutions of iodine, sodium arsenite, potassium permanga- 
nate, sodium thiosulphate. 

Analytical Procedure.* Reduction. The soluble molybdate in amount 
not exceeding an equivalent of 0.5 gram MoOs is placed in a 150-cc. Erlenmeyer 
flask, 20 to 25 cc. of hydrochloric acid (sp.gr. 1.2) added together with 0.2 to 0.6 
gram potassium iodide. A short stemmed-funnel is placed in the neck of the 
flask to prevent mechanical loss during the boiling. The volume of the solution 
should be about 60 cc. The solution is boiled until the volume is reduced to 
exactly 25 cc. as determined by a mark on the flask. The residue is diluted 
immediately to a volume of 125 cc. and cooled. Either process A or B may 
now be followed. 

A. Reozidation by Standard Iodine. A solution of tartaric acid, equiv- 
alent to 1 gram of the solid, is now added, and the free acid nearly neutralized 
with sodium hydroxide solution (litmus or methyl orange indicator) and finally 
neutralized with sodium acid carbonate, NaHCOs, added in excess. A measured 
amount of N/10 iodine is now run in. The solution is set aside in a dark closet 
for two hours, in order to cause complete oxidation, as the reaction is slow. The 
excess iodine is now titrated with N/10 sodium arsenite. 

One cc. N/10 iodine = .0144 gram M0O3 = .0096 gram Mo." 

On long standing a small amount of iodate is apt to form. This is determined 
by making acid with dilute HCl and titrating with N/10 sodium thiosulphate. 

B, Reozidation of the Residue by Standard Permanganate. To the 
reduced solution about 0.5 gram of manganese sulphate in solution is added, 
followed by a measured amount of N/10 permanganate solution, added from a 
burette until the characteristic pink color appears. A measured amount of 
standard N/10 sodium arsenite, equivalent to the pennanganate is then run in 
and about 3 grams of tartaric acid added. The acid is neutralized by acid sodium 
or potassium carbonate, the stopper and the sides of the flask rinsed into the main 
solution. The residual arsenite is now titrated by N/10 iodine, using starch 
indicator. 

Notes. Tartaric acid prevents precipitation during the subsequent neutraliza- 
tion with NaHCOa. A and Ji. 

The addition of manganese salt in B is to i:>revent the liberation of free chlorine 
by the action of KMn()4 on HCl. 

In addition to the oxidation of the lower oxides to molybdic acid, ])otassiuni per- 
manganate added in B liberates free iodine from HI, it ])ro(iuccs iodic acid, and forms 
the higher oxides of manganese. The standard arsenite, on the other hand, converts 
.ree iodine and the iodate to HI and reduces the higher oxides of manganese. 

* F. A. Gooch and Charlotte Fairbanks, Am. Jour. Sc. (4), 2, 160. 
*F. A. Gooch and O. S. Puhnan, Jr. Am. Jour. Sc. (4), 12, 449. 



MOLYBDENUM 



Estimation by Reduction with Jones Reductor and Oxidation by 
Standard Permanganate Solution 

Principle. The procedure depends upon the reduction of molybdic acid 
to MoiOi by passing ita solution through a column of amalgamated zinc into a 
solution of ferric alum, and subsequent oxidation to MoOi by standard potas- 
sium permanganate solution. 

Reactiona. 2MoOt+3Zn =Mo,Oi+3ZnO. 

5Mo,0,+6KMnO.-|-9H^O, = 10MoO,-l-3K,SO.-l-6MnSO,+9H,O. 

Reagents. Potassium permangaTtale approximately N/10 standardized 
against a standard molybdic acid solution. 
10% solution o! ferric alum. 
2.5% solution of sulphuric acid. 
Apparatus. Jones Reductor. 
£ = reductor tube 50 cm. long, 2 cm. inside 
diameter. Smaller tube jirolongation length 
20 cm. inside diameter 0.5 cm. 
Zn>' column of linc 40 cm. long. ZrtahotS mesh 
to Bq.cm.; 
/■= receiving flask; 

P = pre8sure regulator with gauge, set to give 
presBm* in receiving flask of less than 20 cm. 

(?=platinum cone or gauze with mat of fine glass 
wool 2 cm. thick; 
The line in reductor should be protected from 
the air by covering with water, stop-cock S being 
cloeed when not in use. 

Procedure. The receiving flask of the 
Jones reductor, Fig. 49, Ls charged with about 
30 cc. of 10% ferric alum and 4 cc. of phosphoric 
acid.' Through the40-om. column of amalga- 
mated zinc in the reductor are passed in suc- 
cessbn 100 cc. of hot dilute sulphuric acid 
(2.5% sol), the molybdic a::id in the form of 
ammonium iiiolybdate dissolved in 10 cc. of , 
water and acidified with 100 cc. of hot dilute 
Buljihuric acid followed by 200 cc. more of the 
hot dilute sulphuric acid and 100 cc. of hot water. The reduced green molybdic 
acid upon commg in contact with the ferric alum solution produces a bright red 
color. 

The hot solution is titrated with N/10 KMnOt solution. 







9 — Jones Reductor. 



•C. Reinhsrdt, Chem. Ztg., 13, 33. 



282 MOLYBDENUM 

Method for Determining Molybdenum and Vanadium in a 

Mixture of their Acids 

Principle of the Method. The procedure depends upon the fact that 
vanadic acid alone is reduced by SOi^ in a sulphuric acid solution, whereas both 
vanadic and molybdic acids are reduced by amalgamated zinc, in each case the 
reducing agents forming definite lower oxides which are readily oxidized to 
definite higher oxides by KMn04. 

Reactions. 

SOt Reduction: 

1. V,06+S02 = V,04+S0a. (No action on MoO,.) 

Zn Reduction : 

2. V206+3Zn=V202+3ZnO. 

3. 2MoO,+3Zn-Mo20,-f3ZnO. 

KMnO* Oxidation: 

4. 5V204-h2KMn04+3H2S04 =5V205+K2S044-MnS04+3H20. 

5. 5V202-f 6KMn04+9H2S04 =oV20&+3K2S04-f 6MnS04-f 9H2O. 

6. 5Mo20,-f 6KMn04-f 9H2SO4 = 10MoO,-f 3K2S04-f 6MnS04+9H20. 

From the reactions " 4 " and " 5 " it is seen that three times the amount 
of KMn04 is required to oxidize V2O2 to VzOs as is required in the case of V2O4, 
hence — total cc. KMn04 required in oxidation of the zinc-reduced oxides minus 
three times the cc. KMn04 required in oxidizing the tctroxide of vanadium 
formed by the sulphur dioxide reduction ==cc. KMn04 required to oxidize M02OS 
to MoOj. From these data molybdenum and vanadium may readily be calculated. 

Method of Procedure. A. Vanadic Acid. The solution containing the 
vanadic and molybdic acids in a 250- to 300-cc. Erlenmeyer flask, is diluted to 
75 cc. acidified with 2 to 3 cc. of strong sulphuric acid, heated to boiling and the 
vanadic acid reduced by a current of SO2 passed into the solution until the clear 
blue color indicates the complete reduction of the vanadic acid to V2O4. The 
boihng is now continued and CO2 passed into the flask to expel the last trace of S()2. 

Standard N/10 KMn04 is now run into the reduced solution to the character- 
istic faint pink. From reaction " 4,'^ vanadic acid may be calculated. 

One cc. N/10 KMnO4 = .0182 gram V2O6 = .0051 gram vanadium. 

B. Molybdic Acid. The reduction by Jones' reductor, and titration of the 
combined acids reduced by amalgamated zinc with N/10 potassium permanganate 
solution, is carried out exactly as described in the determination of molybdic 
acid alone. In this case 50 cc. of 10% ferric alum and 8 cc. of the phosphoric 
acid is placed in the recei\nng flask. 

Calculation. Total permanganate titration in B minus three times the titration 
in A gives the permanganate recjuired to oxidize M02O3 to M0O3. From equation 
6 the molybdic acid may now be calculated. 

One cc. N/10 K\rn04=^-r — gram M()()3= gram molybdenum. 

* Reduction of vanadium by 8(^2 in presence of molybdenum, Graham Edgar, 
Am. Jour. So., (4) 26, 332. No reduction of MoOj when 0.4 gram is ])resent with o cc. 
H2SO4 in 25 cc. volume. 

For theoretical considerations and data on accuracy of method see " Methods 
in Chemical Analysis," F. A. Gooch. 



NICKEL 

W. L. Saveix 
NUat.wt. 58.68; sp.gr. 8.6-8.9; m.p. 1452''C; 6.p.; oxides, NIO, NIsOs^ NIsOi. 

DETECTION 

After bringing the sample into solution by one of the methods described under 
Preparation and Solution of the Sample, silica is removed, if present, in the usual 
manner, by evaporating the solution to dryness in the presence of an excess of 
hydrochloric acid, dissolving the residue and boiling with hydrochloric acid and 
filtering off the silica. 

Hydrogen sulphide is then passed through the solution to remove the elements 
precipitated by this reagent. The filtrate from this precipitation is then boiled 
to expel the excess of hydrogen sulphide and a little nitric acid added to oxidize 
any ferrous iron to the ferric state. (See page 285, Separations.) Ammonium 
hydroxide is then added to precipitate iron, aluminum and chromium. Cobalt, 
nickel, manganese and zinc are precipitated from the filtrate by adding a solution 
of colorless anmionium sulphide or by passing hydrogen sulphide through the 
ammoniacal solution. Manganese and zinc are separated from the precipitate 
by washing with cold hydrochloric acid of about 1.035 sp.gr. A small quantity 
of the precipitate is fused with borax in the loop of a clean platinum wire. A 
green color in the cool bead indicates nickel. Fairly small quantities of cobalt 
interfere with this test, so if the bead is colored blue it will be necessary to make 
further tests for nickel. 

Dimethylglyoxime will precipitate nickel as oxime from an acetic acid solution 
containing sodium acetate and in this manner separate it from cobalt, manganese 
and zinc. After precipitating iron, aluminum and chromium and filtering them 
off, the solution is slightly acidified with hydrochloric acid, then is neutralized 
with sodium hydroxide, and acidified with acetic acid. A solution of dimethyl- 
glyoxime is added, when nickel, if present, will be precipitated as a flocculent red 
precipitate. 

Nickel may be detected in the presence of cobalt by adding a solution of 
sodium hydroxide to the solution of cobalt and nickel until a slight precipitate 
is formed, then somewhat more potassium cyanide than is necessary to redissolve 
the precipitate and finally two volumes of bromine water. Warm gently and 
allow to stand for some time. If a precipitate of nickel hydroxides separates, 
filter, wash and test with the borax bead. 

Nickel may also be detected in the presence of cobalt by precipitating the 
cobalt as nitrite, as described in the chapter on cobalt, and then precipitating the 
nickel as hydroxide with sodium hydroxide and bromine water and testing the 
precipitate with the borax bead. 

Alpha henzildioxime added to an ammoniacal solution of nickel precipitates 

an intensely red salt having the composition C28H22N404Ni. This precipitate 

is very voluminous. Silver, magi>esium, chromium, manganese and zinc do not 

interfere with this reaction. 

283 



284 NICKEL 



ESTIMATION 

The determination of nickel is required, principally, in the analysis of ores, 
metallic nickel and its alloys, but is also required in the analysis of metallic 
cobalt and cobalt products as well as in a host of miscellaneous m^aterials. 

In the majority of cases the results of a nickel determination are calculated 
in terms of metallic nickel. Even in the determination of nickel in nickel-plating 
solution the results are calculated in terms of metallic nickel since this is the 
factor by which the solutions are controlled. 

Preparation and Solution of the Sample 

The materials in which nickel occurs ordinarily, may, in general, be brought 
into solution by treatment with acids, but in the case of some refractory ores and 
alloys, a fusion is required first to make the acid treatment effective. When 
treating ores containing sulphides or arsenides a strong oxidizing treatment is 
necessary to break up these compounds. Metallic nickel may be dissolved easily 
in nitric acid, more slowly in hydrochloric acid and still more slowly by sulphuric. 
Nickel alloys may be dissolved in a mixture of hydrochloric acid and nitric acid. 

General Procedure for Ores. One gram of the finely powdered ore is weighed 
into a porcelain dish and mixed intimately with 3 grams of powdered potassium 
chlorate. The dish is covered with a watch-glass and 40 cc. concentrated nitric 
acid added slowly. The dish is allowed to stand in a cool place for a few minutes, 
then placed on a water bath and digested until the sample is completely decom- 
posed, stirring the mixture frequently with a glass stirring rod, and adding a little 
potassium chlorate from time to time until the decomposition is complete. The 
watch-glass is then removed and any particles that may have spattered on it 
are washed back into the dish and the evaporation continued to dryness. This 
eva[X)ration to dryness is repeated with the addition of 10 cc. of concentrated 
hydrochloric acid, and the silica dehydrated by heating for an hour or more in an 
air oven at 1 10° C. The dry residue is moistened with concentrated hydrochloric 
acid and the sides of the dish washed do\ni with hot water, the mixture heated 
to boiling and allowed to boil for a few minutes, then withdrawn from the heat 
and filtered, hot, after the insoluble matter has settled. 

Treat the filtrate for the removal of interfering elements as directed under 
Separations. 

Fusion Method. The above method is used where it is desired to determine 
insoluble matter or " gangue." As a method of bringing the nickel in the sample 
into solution it is quite satisfactory and when the insoluble matter bums to a 
pure white ash the ignited residue may be weighed as silica, but in some cases 
this method does not give sufficient information regarding the composition of 
the gangue. 

If it is necessary to make a complete analysis it is usually better to fuse the 
sample with the sodium and potassium carbonate mixture containing a little 
potassium nitrate and then treat in the usual manner to determine sili(?a. 

Potassium Bisulphate Fusion. In the treatment of nickel and cobalt oxides 
these are ground to a fine powder and a representative sample of 1 gram is fused 
with 10 grams of potassium bisulphate. This may be done in a porcelain or 
8'Hca crucible or dish. The melt is extracted with water and the silica filtered off. 

A small casserole has been found to be very useful for this fusion. 



NICKEL 285 

Solution of Metallic Nickel and Its Alloys. From 1 to 5 grams of the well- 
mixed drillings are treated with a minimum quantity of nitric acid and 20 cc. 
1 : 1 sulphuric acid added and the solution evaporated to fumes of sulphur tri- 
oxidc. Allow the fuming to continue for ten minutes. Dilute carefully with a 
little water and filter ofif the insoluble. Continue as directed in the following 
detailed analyses. 

It may be necessary to use a mixture of nitric and hydrochloric acids to bring 
certain aUojrs into solution, after which the procedure is the same as above. 

SEPARATIONS 

Separation of the Ammonium Sulphide Group, Containing Nickel from 
the Hydrogen Sulphide Group. Mercury, Lead, Bismuth, Copper, Cadmium, 
Arsenic, Antimony, Tin, Gold, Molybdenum, etc. 

The hydrogen sulphide group elements are precipitated from an acid solution 
(HCl) by HaS, and removed by filtration, nickel, etc., passing into the filtrate. 

Separation of the Ammoniimi Sulphide Group from the Alkaline Earths 
and Alkalies. Nickel is precipitated with other members of the group by 
passing H2S into its ammoniacal solution, or by adding (NH4),S solution. The 
alkaline earths and alkalies are not precipitated. 

Separation of Nickel from Cobalt This procedure can be carried out in 
exactly the same manner as the method given for the determination of nickel 
by precipitation of nickel with dimethylglyoxime, since cobalt is soluble as oxime. 
In case more cobalt is present than nickel a larger excess of the reagent must be 
used. The excess of acid is best neutralized with ammonium hydroxide. If 
both metals are to be determined, cobalt may be determined electrolytically 
in the filtrate. 

An alternate method is to determine the cobalt and nickel as oxides, or metal 
by electrolysis, together. The oxides, or plate, are dissolved in nitric acid and 
the nickel determined in the solution, cobalt being found by difference. 

For other methods see Separation of Cobalt from Nickel, under Cobalt, page 142. 

Separation of Nickel from Manganese. Nickel is precipitated by dimethyl- 
glyoxime from an acetic acid solution containing sodium acetate, manganese 
being determined in the filtrate. 

Separation of Nickel from Zinc. Zinc does not interfere in the dimethyl- 
glyoxime precipitation of nickel when ammonium salts are present. It is advis- 
able to precipitate the nickel in a dilute acetic acid solution, thus avoiding the 
addition of a large amount of ammonium salts as would be necessary if the pre- 
cipitation took place in an ammoniacal solution. Zinc readily remains in solution, 
and may be determined in the filtrate from the nickel oxime precipitate. The 
following procedure is recommended: 

The solution containing the two metals is neutralized with ammonium hydrox- 
ide and then made just slightly acid with acetic acid and sodium acetate added. 
Dimethylglyoxime solution is now added to the solution, which is nearly boiling, 
and the procedure given for the determination of nickel by this reagent is followed. 

Separation of Nickel from Iron. Nickel cannot be separated satisfactorily 
from iron by precipitating the latter 'N^ith anunonium hydroxide, as some of the 
nickel is invariably occluded by the ferric hydroxide precipitate. Two modi- 
fications of the oxime method may be used. 

(1) The iron, if present as a ferric salt, is converted into a complex salt by 



286 NICKEL 

adding from 1 to 2 grams of tartaric acid, and the solution diluted to 200 or 300 
CO., boiled and the nickel precipitated as the oxime in an ammoniacal solution 
by the prescribed method. Iron fonns no oxime under these conditions. 

The iron may be precipitated from this filtrate by colorless anamonium sul- 
phide and the sulphide converted to ferric oxide (P'ejOj) by ignition. 

(2) Ferric iron is reduced to the ferrous condition by warming with sulphurous 
acid, in a nearly neutral solution. If the original solution has an excess of acid, 
it is treated with a solution of sodium hydroxide until a permanent precipitate 
is formed. This is dissolved with a few drops of hydrochloric acid and the iron 
reduced by adding from 6 to 10 cc. of a saturated solution of sulphur dioxide or 
by passing dioxide through the solution. The solution is diluted to 200 or 300 cc. 
and the solution of dimethylglyoxime added in slight excess, followed by sodium 
acetate until a permanent precipitate of nickel oxime is formed. After adding 2 
grams more of sodium acetate the solution is filtered immediately. The iron 
is precipitated from the filtrate by oxidizing with bromine water and adding 
ammonium hydroxide to precipitate the basic acetate of iron. 

Procedure (1) is suitable for the determination of nickel in iron and steel. 

Separation of Nickel from Aluminum. This method is the same as pro- 
cedure (1) given above. 

Separation of Nickel from Chromium. This separation cannot be carried out 
in an acetic acid solution. From 1 to 2 grams of tartaric acid are added and from 
5 to 10 cc. of a 10% ammonium chloride solution, subsequently. The solution 
is made ammoniacal, but no precipitate should form. If the solution becomes 
cloudy, it is acidified with hydrochloric acid and additional anunonium chloride 
added and again made ammoniacal and the nickel precipitated as oxime accord- 
ing to directions given from this precipitation. 



GRAVIMETRIC METHODS FOR THE DETERMINATION 

OF NICKEL 

Precipitation of Nickel by Alpha Benzildioxime 

The alcoholic solution of alpha benzildioxime gives an intensely red precipi- 
tate of C»Hi2N404Ni, when added to ammoniacal solutions containing nickel. 
The reaction is more characteristic for nickel than is that with dimethylgly- 
oxime and is more delicate. In a volume of 5 cc. (according to F. H. Atack), 
1 part of nickel in 2,000,000 parts of water may be detected. In the presence of 
100 times as much as cobalt only a faint yellow color is produced by the cobalt. 
One port of nickel per million of water will cause precipitation with the compound, 
whereas no precipitate is formed with dimethylglyoxime under the same condi- 
tions. With glyoxime iron produces a pink color, with alpha l)enzildioxime 
ferrous salts give a faint violet color, hence do not interfere in the detection of 
nickel. Silver, magnesimn, chromium, manganese, and zinc do not interfere. 
Since the nickel precipitate with this reagent is exceedingly voluminous it is 
advisable to have not more than 0.025 gram of nickel in the solution in which the 
nickel is being determined. The method is adapted to the detection and 
determination of minute traces of the element up to small amounts of less than 
10% nickel. 



NICKEL 287 

Reagent, Alpha Benzildioxime. This may be prepared by boiling 10 grams of 
benzil (uot necessarily pure) with 8 to 10 grams of hydroxylamine hydrochloride 
in methyl alcohol solution. After boilmg for three hours the precipitate is 
filtered off and dried, washed with hot water and then with a small amount of 
50% alcohol, and dried. This dried precipitate consists of pure benzildioxiir.e 
(m.p. 237** C.). A further yield may be obtained by boiling the filtrate with 
hydroxylamine hydrochloride. The reagent is prepared by dissolving 0.2 gram 
of the salt per liter of alcohol to which is added ammonium hydroxide to make 
5% solution, sp.gr. 0.96 (50 cc. per liter). 

Procedtire. A slight excess of the warmed solution of the above reagent is 
stirred into the ammoniacal solution containing nickel and the whole heated 
on the water bath for a few moments to coagulate the precipitate. Quantitative 
precipitation is complete after one minute. The liquid is filtered through a 
Gooch crucible, with suction, or onto a filter paper, for which a counterpoise 
has been selected. The counterpoise paper is treated in exactly the same manner 
as the one containing the precipitate. The precipitate is washed with 50% 
alcohol, followed by hot water, and is then dried at 1 10° C. In weighing the 
precipitate the counterpoise filter is placed in the weight pan of the balance. The 
precipitate contains 10.93% nickel. Weight of C28H,2N404NiX 0.01093 =Ni. 

Notes. Acetone may be used instead of alcohol as a solvent of the reagent. The 
compound is more soluble in acetone than in alcohol. 

The precipitate does not pass through the filter as does the compound with dimethyl- 
glyoxime. 

The method is affected by the presence of nitrates, hence these must be removed 
by evaporation of the solution with sulphuric acid to fumes, before the addition of 
the reagent to the nickel solution. 

In the presence of cobalt an excess of the reagent must be used, as in the case of 
the dimethylglyoxin.e precipitation. 

In the presence of iron and chromium Rochelle salt, sodium citrate or tartiiric 
acid are added to prevent precipitation of the hydroxides of these metals upon making 
the solution alkaline. 

In the presence of manganese a fairly large excess of the reagent is required, the 
solution being slightly acid with acetic acid. 

Zinc and magnesium arc kept in solution by addition of ammonium chloride. 

Large amounts of copper must be removed by precipitating with hydrogen sul- 
phide before addition of the reagent. 

The nickel salt with the reagent forms an extremely voluminous precipitate so that 
a concentration of 0.09 gram of nickel per 250 cc. is as high as is desirable. The 
process is applicable to the determination of nickel in the filtrate obtained in the 
separation of zinc after the removal of the hydrogen sulphide, formic acid, etc. 

Method by F. W. Atack, The Analyst, 38, 448, 318. Cockbum, Gardiner and 
Black, Analyst, 38, 439, 443. 

Precipitation of Nickel by Dimethylglyoxime 

Preliminary Considerations. This method has been demonstrated by O. 
Bninck to be the most accurate and expeditious procedure known for nickel.* 
By this method 1 part of nickel may be detected when mixed with 5000 parts of 
cobalt or 1 part of nickel may be detected in 400,000 parts of water. The 
nickel precipitate with this reagent is almost completely insoluble in water and is 
only very slightly soluble in acetic acid, but is easily decomposed by strongly 
dissociated acids, so that the precipitation m incomplete in neutral solutions of 
nickel chloride, sulphate or nitrate. If, however, the free acid formed is neutral- 

iZeit. f. ang. Chem., 20, 1844. 



288 NICKEL 

ized with sodium, potassium or ammonium hydroxides or by addition of the ace- 
tate salts of these bases, nickel will be completely precipitated, not even a trace 
being found in the filtrate. 

" The quantitative determination of nickel in the presence of other metals 
is a simple operation. The nickel should be in the form of a convenient salt. 

" The concentration of the solution does not matter; the precipitation can 
take place either in a solution of the greatest concentration, or in a very dilute 
solution. The reaction is not hindered by the presence of ammonium salts." 

Iron, aluminum, chromium, cobalt, manganese and zinc do not interfere. 
Theoretically 4 parts of dimethylglyoxime, added as a 1% alcoholic solution, 
are necessary; a certain excess does no harm provided the alcohol volume does 
not exceed more than half that of the water solution containing the nickel salt, 
as alcohol has a solvent action on the oxime. The compound is very stable and 
volatilizes undecomposed at 250° C. 

An excess of ammonium hydroxide is also to be avoided in the solution in which 
the precipitation takes place. 

It has been observed that the precipitate of nickel with dimethylglyoxime 
may be safely ignited to the oxide NiO without loss, if the filter is first care- 
fully charred without allowing it to take fire, then gradually heated to redness. 

Procedure. Such an amount of the sample should be taken that the nickel 
be not over 0.1 gram, as glyoxime of nickel is very voluminous and a larger amount 
would be difficult to filter. If cobalt is present it should not exceed 0.1 gram 
in the sample taken.* 

If hydrogen sulphide has been used to precipitate members of the second 
group, it is expelled by boiling the acid solution and the volume brought to 250 cc. 

One or 2 grams of tartaric acid are added to prevent the precipitation of 
the hydroxides of iron, aluminum and chromium by ammonium hydroxide 
(this treatment is omitted if these are absent), and 5 to 10 cc. of a 10% solu- 
tion of ammonium chloride added to keep zinc and manganese in solution, should 
they be present. Ammonium hydroxide is now added until the solution is slightly 
alkaline. If a precipitate forms, ammonium chloride is added to clear the 
solution, followed by ammonium hydroxide to neutralize the acid. The solu- 
tion should remain clear after this treatment, otherwise the ammonium chloride 
is added in solution or as salt until the solution of the sample will remain clear. 
It is then heated to nearly boiling and the alcoholic solution of dimethylglyoxime 
added until the reagent is approximately seven times, by weight, the weight of 
nickel present. Ammonium hydroxide is now added until the solution has a dis- 
tinct odor of this reagent. The precipitation of the scarlet red nickel salt is hast- 
ened by stirring. It is advisable to place the mixture on the steam bath for 
fifteen to twenty minutes to allow the reaction to go to completion before filter- 
ing. The precipitate Is filtered off, into a platinum sponge Gooch crucible, some- 
times known as a Neubauer Gooch crucible. (Other forms of Gooch crucible 
are used for this purpose, but the Neubauer crucible has been found to be most 
satisfactory.) The precipitate is dried for about two hours at 110 to 120° C. 
and weighed as C8HHN404Ni, which contains 20.32% Ni- 

Weight of precipitate multiplied by 0.2032 = weight of nickel. 

* If the sample contains more than 0.1 gram of col)alt, a large excess of ammo- 
nium hydroxide and dimethylglyoxime is necessary to prevent its precipitation, 
hence it is advisable to take such weights of samples that the cobalt content will be less 
than this weight. 



NICKEL 289 

In plftoe of a Gooch crucible a tared filter paper mav be used. It must be remem- 
bered, however, that a blank filter paper of tnc same kind as used for the precipitate" 
must be used as a counterbalance, after treating in exactly the same manner as the 
(me containinii^ the precipitate. This is necessary because it has been found that filter 
paper loees weight ouring washing and drying. 

Precipitation of Nickel by Electrolysis ^ 

This precipitation Is conducted in exactly the same manner as the one 
described under Cobalt for the Precipitation of Cobalt by Electrolysis, and 
requires that the same precautions be exercised in the practice of the method. 

In the presence of cobalt the two elements may l)e determined together 
by electrolysis as described below and the deiK)sit^l metal redissolved and the 
two elements separated by one of the methods given under Cobalt or Nickel. 

Procediire. After the sample has been brought into solution by one of the 
methods outlined under Preparation and Solution of the Sample, the solution 
is evaporated with 20 cc. of 1 : 1 sulphuric acid for every gram of metal in the 
sample. The evaporation is continued until the solution has fumed strongly 
for ten minutes. Cool carefully and dilute with 20 cc. of water. Heat the solu- 
tion to nearly boiling and pass hydrogen sulphide for one hour to precipitate 
members of the second group. This long treatment is necessary to insure com- 
plete precipitation of arsenic. Filter and boil to expel hydrogen sulphide. Add 
5 cc. nitric acid to insure oxidation of iron compounds to the ferric state and add 
ammonium hydroxide until just slightly alkaline. Filter off the ferric hydroxide 
and wash with water containing a small quantity of ammonium hydroxide. To 
recover occluded nickel dissolve the precipitate in hydrochloric acid and repre- 
cipitate the iron with addition of a little hydrogen peroxide. Combine the 
filtrates. Eva[X)rate to about 250 cc. and add 50 cc. of strong ammonium 
hydroxide and electrolyze as descril^ed under Cobalt, page 144. 

The increase in weight of the electrode is the weight of cobalt and nickel 
in the sample. The percentage of cobalt and nickel in the sample is found by 
multiplying the increase in weight of the electrode by 100 and dividing by the 
weight of the sample. 

Note. The deposition of cobalt and nickel by the alx)vc method has l)een found 
to be the most accurate of the elect rolvtic methods. In the solutions containing the 
organic acids there is always more or less carbide deix)sited on the cathode with the 
metal. This causes high results. 

Nickel in Metallic Nickel 

This determination may be made in the manner described under Precipitation 
of Nickel by Electrolysis, separating cobalt before or after the electrolysis or by 
the method described under Prectipitation of Nickel by Dimethylglyoxime. The 
latter method is recommended. 

Nickel in Cobalt and Cobalt Oxide 

The dimethylglyoxime precipitation is used in combination with the eleo- 
trol3rtic precipitation. See chapter on Cobalt. 

» W. J. Marsh, J. Phys. Chem., 18, 705-16, 1914. 



290 NICKEL 

VOLUMETRIC DETERMINATION OF NICKEL 
Determination of Nickel in Alloys 

This method, as described by S. W. Parr and J. M. Lindgren,* consists of 
H modification of the dimethylgloxime method. The precipitation takes place 
in the usual manner and the precipitate is dissolved in sulphuric acid and the 
excess titrated with a standard solution of potassium hydroxide. 

procedure. The alloy is dissolved in nitric or hydrochloric acids and if iron, 
aluminum or chromium are present twice their weight of tartaric acid is added 
to prevent their precipitation. If chromium is present anunonium chloride is 
also added. If manganese or zinc is present hydrochloric acid should be used and 
paost of the free acid evaporated. Add a few cc. of hydrogen peroxide to oxidize 
any ferrous iron to the ferric state. Dilute to 300 or 400 cc. and neutralize the 
free acid by sodium acetate. Heat the solution to nearly boiling and add five times 
as much dimethylglyoxime, in 1% alcoholic solution, as the nickel present. Then 
completely neutralize with ammonium hydroxide, using a very slight excess (or 
the solution may be neutralized with sodium acetate). Heat until all the nickel 
is precipitated. Filter and wash. Place the precipitate and filter in a beaker, 
ftdd an excess of 0.05N sulphuric acid, dilute to 200 cc, heat until solution is 
complete and titrate back with O.IN potassium hydroxide solution, taking the 
first faint yellowish tinge as the end-point. The solutions are standardized 
against pure nickel. 

Note. Cobalt should not exceed 0.1 gram per 100 cc. and an excess should be 
used of the dimethylglyoxime. 

Nickel in Nickel-plating Solutions 

In most cases it is quite unnecessary to separate the cobalt from the nickel 
in making this determination and, as the principal impurity is usually iron, the 
best practice is to follow the method given under Precipitation of Cobalt by 
Electrolysis, page 144. 

If chlorides or organic matter are present in the solution the preparation 
of the solution for electrolysis is accomplished in the following manner: 

From the well-stirred solution in the plating tank, withdraw about 200 cc. 
and place in a small beaker. Prepare a 100-cc. burette by thoroughly clean- 
ing it with the sulphuric acid and potassium bichromate mixture and distilled 
water. Wash finally with a few cc. of the nickel solution and fill the burette with 
the solution from the plating tank. 

Run 66.7 cc. into an eva[X)rating dish and add 2 cc. 1 : 1 sulphuric acid. 
Evaporate to fumes of sulphur trioxide and allow to fume strongly for ten min- 
utes. Dissolve in a little water. Dilute to 200 cc. carefully, neutralize with a 
solution of ammonium hydroxide and add 50 cc. of strong anunonium hydroxide 
and electrolyze. (See Precipitation of Cobalt by Electrolysis.) 

The increase in weight of the cathode in grains multiplied by 2 gives the 
weight in ounces of nickel in one United States gallon of the plating solution. 

* S. W. Parr and J. M. Lindgrcn, Trans. Am. Brass Founders' Assoc, 6, 120-9. 



NITROGEN 

Wilfred W. Scott 

Element. Nt, at.wt.l^.Ol; D. (air) 0.9674; m.p. -210"; 6.p. -195.5'' C; 
oxides, NsO, NsOz, NzO,, l^tOu N^Oi. 

Ammonia. NHs, m.w. 17.03; D. (air) 0.5971; sp.gr. liquid 0.6234; m.p. 
-77.3**; b.p. -38.5** C. Crit. temp. 130**; liquid at 0** with 4.2 atmospheres 
pressure. Commercial 28% NH39 sp.gr. 0.90. 

Nitric Acid. HNO„ m.ir., 63.02; «p.^r. 1.53; m.p. -41.3; 6.p. 86**C. 
Boiling-point of commercial 95% acid is a little above 86°, but gradually 
rises to 126** and the strength of acid falls to 68.9%, sp.gr. is then 1.42. 
The acid now remains constant, the distillate being of the same strength. 

DETECTION 

Element. Organic Nitrogen. Organic matter is decomposed by heating 
in a Kjeldahl flask with concentrated sulphuric acid as described under prepara- 
tion and solution of the sample. Ammonia may now be liberated from the sul- 
phate and so detected. 

Nitrogen in Gas. Recognized by its inertness towards the reagents used 
in gas analysis. The element may be recognized by means of the spectroscope. 

Ammonia. Free ammonia is readily recognized by its characteristic odor. 
A glass rod dipped in hydrochloric acid and held in fumes of ammonia produces 
a white cloud of ammonium chloride, NH4CI. 

Moist red litmus paper is turned blue by ammonia. Upon heating the paper 
the red color is restored, upon volatilization of ammonia (distinction from fixed 
alkalies). 

Nessler*s Test.^ Nessler's reagent added to a solution containing ammonia, 
combined or free, produces a brown precipitate, NHgjI • H2O. If the ammoniacal 
solution is sufficiently dilute a yellow or reddish-brown color is produced, accord- 
ing to the amount of ammonia present. The reaction is used in determining 
ammonia in water. 

Salts of ammonia are decomposed by heating their solutions with a strong 
base such as the hydroxides of the fixed alkalies or the alkaline earths. The 
odor of ammonia may now be detected. 

Nitric Acid. Ferrous Sulphate Test. About 1 to 2 cc. of the concentrated 
solution of the substance is added to 15 to 20 cc. of strong sulphuric acid in a 
test-tube. After cooling the mixture, the test-tube is inclined and an equal 
volume of a saturated solution of ferrous sulphate is allowed to flow slowly down 
over the surface of the acid. The tube is now held upright and gently tapped. 
In the presence of nitric acid a brown ring forms at the junction of the two 
solutions. 

^The reagent is made by dissolving 20 grams of potassium iodide in 50 cc. of water, 
adding 32 grams of mercimc iodide and diluting to 200 cc. To this is added a solution 
of potassium hydroxide — 134 grams KOH per 260 cc. H2O. 

291 



292 



NITROGEN 



The test for nitrate may be made according to the quantitative procedure 
given for determining of nitric acid (see later). It should be remembered that 
ferrous sulphate should be present in excess, otherwise the brown color is 
destroyed by the free nitric acid. Traces of nitric acid in sulphuric produce 
a pink color with the sulphuric acid solution of ferrous sulphate. (See Deter- 
mination of Nitric Acid — Ferrous Sulphate Method.) 

Ferro- and ferricyanides, chlorates, bromides and bromates, iodides and 
iodates, chromates and permanganates interfere. 

Diphenylamine Tests for Nitrates. (C»H6)2NH dissolved in sulphuric acid 
is added to 2 or 3 cc. of the substance in solution on a watch-glass. Upon gently 
warming a blue color is produced in presence of nitrates. Nitric acid in sul- 
phuric acid is detected by placing a crystal of diphenylamine in 3 or 4 cc. of the 
acid and gently warming. CI', Cr, Br^ T, Mn^, Cr^, Se"^, Fe'" interfere. 

Copper placed in a solution containing nitric acid Uberates brown fumes. 

Phenolsulphonic Acid Test. See chapter on Water Analysis. 

Detection of Nitrous Acid. Acetic Acid Test. Acetic acid added to a 
nitrite in a test-tube (inclined as directed in the nitric acid test with ferrous 
sulphate), produces a brown ring. Nitrates do not give this. If potassium iodide 
is present in the solution, free iodine is liberated. The free iodine is absorbed 
by chloroform, carbon tetrachloride or disulphide, these reagents being colored 
pink. Starch solution is colored blue. 

Nitrous acid reduces iodic acid to iodine. The iodine is then detected with 
starch, or by carbon disulphide, or carbon tetrachloride. 

Potassium Permanganate Test. A solution of the reagent acidified with sul- 
phuric acid is decolorized by nitrous acid or nitrite. The test serves to detect 
nitrous acid in nitric acid. Other reducing substances must be absent. 

ESTIMATION 

Occurrence. Element. Free in air to extent of 78%+ by volume and 
76% - by weight. 

Air weight of 1 liter = 1.293 grams. With oxygen as 32, air =28.95. 

Composition of Air. On the Basis op 1000 Liters of Atmosphere 



Element. 



Nitrogen 

Oxygen 

Argon 

Carbon dioxide, 

Hydrogen 

Neon 

Helium 

Krypton 

Xenon 



Liters per 
1000 1. 



780.3 

209.9 
9.4 
0.3 
0.1 
0.015 
0.0015 
0.00005 
0.000006 



Weight per 
1000 1. grams 



975.80 

299.84 
16.76 
0.59 
0.01 
01339 
0.00027 
0.00018 
0.00003 



Per cent by 
Vol. 



78 1 

21.0 
0.9 
0.04 



Per cent by 
Wt. 



75.47- 

23.19- 

1.296+ 
0.045 



Water-saturated air contains 2.4 grams H^O at —10°; 4.9 grains at 0®; 17.2 
grains at 20** and 55 grams HjO at 40° C. Ordinarily 50 to 70% of this is present. 

Nitrogen is found combined in nature as potassium nitrate (saltpeter), KNOi; 
sodium nitrate (Chili saltpeter), NaNOs, and to a less extent as calcium nitrate. 



NITROGEN 293 

Ca(N0i)2. It occurs in plants and in animals, in the substances proteids, blood, 
muscle, nerve substance, in fossil plants (coal), in guano, ammonia and ammo- 
nium salts. 

Free nitrogen is estimated in the complete analysis of gas mixtures. In 
illuminating gas the other constituents are removed by combustion and absorp- 
tion and the residual gas measured as nitrogen. 

Total nitrogen in organic substances is best determined by decomposition 
of the materials with sulphuric acid as described later, and estimating the nitro- 
gen from the ammonia formed. 

Combined nitrogen in the form of ammonia and nitric acid specially concerns 
the analyst. In the evaluation of fertilizers, feedstufifs, hay, fodders, grain, 
etc., the nitrogen is estimated after conversion to ammonia. Ammonia, nitrates 
and nitrites may be required in an analysis of sewages, water, and soils. Nitric 
acid is determined in Chili saltpeter, in the evaluation of this material for the 
manufacture of nitric acid or a fertilizer, the nitrate being reduced to ammonia 
and thus estimated. 

We will take up a few of the characteristic substances in which nitrogen 
estimations are required, e.g., in organic substances as proteids, in soils and 
fertilizers; in ammonium salts, nitrates, and nitrites, free ammonia in ammonia- 
cal liquors, nitric acid in the evaluation of the commercial acid and in mixed acids. 

In general nitrogen is more accurately and easily measured as ammonia, to 
which form it is converted by reduction methods. Large amounts are determined 
by titration, whereas small amounts are estimated colorimetrically. Nitric 
acid and nitrates may be determined by direct titration by the Ferrous Sulphate 
Method outlined later. The procedure is of value in estimation of nitrates in 
mixed acids. The nitrometer method for determining nitrates (including 
nitrites), and the free acid in mixed acids, is generally used by manufacturers 
of explosives. 

Preparation of the Sample 

It will be recalled that compounds of ammonia and of nitric acid are generally 
soluble in water. All nitrogen compounds, however, are not included. Among 
those which are not readily soluble the following deserve mention: compounds 
of nitrogen in many organic substances; nitrogen bromophosphide, NPBrj; 
nitrogen selenide, NSe; nitrogen sulphide, N4S4; nitrogen pentasulphide, NaSs; 
ammonium antimonate, NH4Sb03*2H20; ammonium iodate, HNJOi (2.6 grams 
per 100 cc. H2O); ammonium chlorplatinate, (NH4)2PtCl« (0.67 gram); ammo- 
nium chloriridate, (NH4)2lrCl« (0.7 gram); ammonium oxalate, (NH4)8Ci04"H20 
(4.2 grams); ammonium phosphomolybdate, (NH4)8P04-12MoOi (0.03 gram); 
nitron nitrate, C2oHi6N4'HN08. 

Organic Substances 

By oxidation of nitrogenous organic substances with concentrated sulphuric 
acid, containing mercuric oxide, or potassium permanganate, the organic matter is 
destroyed and the nitrogen is changed to ammonia, which is held by the sul- 
phuric acid as sulphate. Nitrates are reduced by addition of salicylic acid, 
sine dust, etc., previous to the oxidation process. Practically all the procedures 
are based on the Kjeldahl method of acid digestion. The modification, com- 
monly known as the Kjeldahl-Gunning-Amold Method, is as foUo^Ys: 



294 



NITROGEN 



Method in Absence of Nitrates. Weight of Sample. Fertilizers 0.7 to 
3.5 grams. Soila 7 to 14 gramB. Meat and meat products 2 grams. Milk 5 
grams. The amount of the substance to be takes should be governed by its 
nitrogen content. ■ 

Acid Digestion.) The material is placed in a Kjcldahl flask of about 550 cc. 
capacity. Approximately 0.7 gram of mercuric oxide or an equivalent amount 
of metallic mercury together with 10 
grams of powdered potassium sulphate 
followed by 20 to 30 cc. of concentrated 
sulphuric acid {sp.gr. 1.84) are added. 
The flask is placed in an inclined posi- 
tion, resting in a large circular opening 
of an asbestos board. The flask is heated 
with a small flame until the frothing has 
ceased, (A piece of paraffin may be 
added to prevent extreme frothing.) 
The heat is then raised and the acid 
brought to brisk boihng, the heating 
being contbued until the solution be- 
comes a pale straw color, or practically 
water white. (In case of leather, scrap, 
cheese, milk products, etc., a n'.ore pro- 
longed digestion may be required. With 
a good flame from one-half to one hour 
of acid digestion is generally sufficient 
to completely decompose the material.) 
The flask is now removed from the flame 
and after cooling the solution is diluted 
with about 200 cc. of water and a few 
pieces of granulated zinc added to pre- 
vent " bumping " (50 nig. or so of No. 
SO granulated zinc). The solution is 
now alkalized strongly by addition of a 
mixture of sodium hydnixidc and sodium sulphide solution (about 75 cc. of a 
mixture containing 25 grams of NaOH and 1 gram Na,S). Phenolphthalein 
indicator added to the solution will show when the acid is neutralized. The 
flask is connected by means of a Hopkins distillation tube (Fig. 53) to a con- 
denser and about 150 cc. of the solution distilled into an excess of standard sul- 
phuric acid and the excess of the acid detemiiiied by titration with standard 
sodium hydroxide. (Methyl red indicator.) 

The ammonia niay be absorbed in a saturated solution of boric acid and 
titrated directly with standard acid. (Methyl orange indicator.)' 

One cc. N/10 H^O. =0.001704 gram NH,. 

' Bee data of arproximatc nitrogen content in certain nitrogenous Bubslances, Jour, 
Ind. KnR. Chem., 7, 357, 1913. 

* Fig. 50 shows n compact a]>pnratus with several seta of flasks and condensetSj 
which enable half a doicn or more determ I nations to bo made at one time. 

• L. W. Winkler, Z. aagew. Chem,, 27, 1, 630-2, 1914. E. UemarU. ibid., 87, 1, 
664, 1914. 




Fio. 50. 
Apparatus for Determining Nitrogen. 



NITROGEN 295 

In Presence of Nitrates. The procedure differs from the former in the 
preliminary treatment to reduce the nitrates. The material in the flask is 
treated with a mixture of 30 to 35 cc. of strong sulphuric acid containing 1 gram 
of salicylic acid and the mixture shaken and allowed to stand for five to ten 
minutes with frequent agitation. About 5 grams of sodium thiosulphate are now 
added and the solution heated for five minutes. After cooling, mercuric oxide 
or metallic mercury and potassium sulphate are added, and the solution treated 
as directed above. 

Notes. Mercuric oxide or metallic mercury are added as a catalyzer to assist the 
oxidation of the organic matter. The digestion process is shortened considerably 
by its use. In place of mercuric oxide or the metal, copper sulphate may be used. 
In this case the addition of sodium sulphide is omitted. Copper sulphate acts as an 
indicator in the neutralization of the sample with caustic. 

Potassium sulphide is added to remove the mercury from the solution and prevent 
the formation of mercur-ammouium compounds, which are not completely decomposed 
by sodium hvdroxide. 

A blank determination should be made on the reagents used with sugar as the 
organic substance. 

Soils. Available Nitrates. Five hundred to 1000 grams of the air-dried 
soil is extracted with 1 to 2 liters of water containing 10 to 20 grams of dextrose. 
Fifteen to twenty hours of leaching is sufficient. An aliquot portion is taken 
for analysis. 

Ammonium Salts. The sample is placed in the distillation flask with splash 
bulb as described in the modified Kjeldahl procedure for organic substances, 
and the material decomposed with ammonia-free caustic solution. The ammonia 
is distilled into an excess of standard acid or a saturated solution of boric acid 
(neutral to methyl orange), and the ammonia determined as usual, either by 
titration of the excess of acid, or by direct titration with acid, according to the 
absorbent used. 

Nitrates. The sample, broken down as fine as possible, is dissolved in 
water, decomposed with Devarda alloy and distilled as described by the modi- 
fied Devarda methods given later. 

Nitrites. The material, dissolved in water, is titrated with standard perman- 
ganate solution according to the procedure described later. 

Mixtures of Ammonium Salts, Nitrates, and Nitrites. Ammonia is 
determined by distillation with caustic as usual. The nitrite is titrated with 
permanganate. Total nitrogen is determined by the modified Devarda methods. 
Nitric acid is now estimated by difference, e.g., from the total nitrogen is deducted 
the nitrogen due to ammonia together with the nitrogen of the nitrite and the dif- 
ference calculated to the nitrate desired. The nitrate may be determined in pres- 
ence of nitrite and ammonia by direct titration with ferrous sulphate. The 
detailed procedures may be found under the Volumetric Methods. 

Nitric Acid in Mixed Acid. This is best determined by the ferrous sulphate 
method for nitric acid. The nitrometer method is also excellent. 

SEPARATIONS 

Ammonia. No special separation need be considered in the determination 
of ammonia. The general method has already been mentioned by which 
ammonia is liberated from its salts by a strong base and volatilized by heat. 
This effects a separation from practically all substances. 



296 NITROGEN 

Nitric Acid. The compound may be isolated as the fairly insoluble, cryis- 
talline nitron nitrate, C2oHieN4*HNOi by the following procedure. 

Such an amount of the substance is taken as will contain about 0.1 gram 
nitric acid, and dissolved in about 100 cc. of water with addition of 10 drops 
of dilute sulphuric acid. The solution is heated nearly to boiling and about 
12 cc. of nitron acetate solution added (10 grams of nitron in 100 cc. of 5% acetic 
acid).* The solution is cooled and placed in an ice pack for about two hours, 
and the compound then transferred to a Gooch or Munroe crucible (weighed 
crucible if gravimetric method is to be followed), and after draining, it is washed 
with about 10 to 12 cc. of ice-water added in small portions. The nitrate may 
now be determined gravimetrically by drying the precipitate to constant weight 
at 110** C, 16.53% of the material being due to NO,. 

The base diphenyl-endo-anilo-hydro-triazole (nitron) also precipitates the 
following acids: nitrous, chromic, chloric, perchloric, hydrobromic, hydriodic, 
hydroferro- and hydroferricyanic, oxalic, picric and thiocyanic acids. Hence 
these must be absent from the solution if precipitation of nitric acid is desired for 
quantitative estimation. 

Removal of Nitrous Aeid. Finely powdered hydrazine sulphate is dropped 
into the concentrated solution. (0.2 gram substance per 5 or 6 cc.) 

Chromic acid is reduced by addition of hydrazhie sulphate. 

Hydrobromic acid is decomposed by chlorine water added drop by drop 
to the neutral solution, which is then boiled until the yellow color has dis- 
appeared. 

Hydriodic acid is removed by adding an excess of potassium iodate to 
the neutral solution and boiling until the iodine is expelled. 



PROCEDURES FOR THE DETERMINATION OF COMBINED 

NITROGEN 

Ammonia 

The volumetric procedures for determination of ammonia are preferred to the 
gravimetric on account of their accuracy and general applicability. The fol- 
lowing gravimetric method may occasionally be of use: 

Qravimetric Determination of Ammonia by Precipitation as 

Ammonium Platinochloride, (NH4)2PtCl6 

Ammonia in ammonium chloride may be determined gravimetrically by pre- 
cipitation with chlorplatinic acid. The method is the reciprocal of the one for 
determining platinum. 

Procedure. The aqueous solution of the ammonium salt is treated with an 
excess of chlorplatinic acid and evaporated on the steam bath to dr>'ness. The 
residue is taken up with absolute alcohol, filtered through a weighed Gooch cru- 
cible, and washed with alcohol. The residue may now be dried at 130° C\ and 
weighed as (NH4)2PtCle, or it may be gently ignited in the covered crucible until 

» M. Busch, Ber., S8, 861 (1905), Treadwell and Hall, " Analytical Chemistry." 
* Keep nitron reagent in a dark-colored bottle. 



NITROGEN 297 

ammonium chloride has been largely expelled and then more strongly with free 
access of air. The residue of metallic platinum is weighed. If the ignition 
method is to be followed, the ammonium platinic chloride may be filtered into a 
small filter, the paper with the washed precipitate placed in a porcelain crucible 
and then gently heated until the paper is charred (crucible being covered) and then 
more strongly with free access of air until the carbon has been destroyed. 

Factors.1 (NHO^PtCU X 0.2400 =NH4C1, or 0.08095 =NH4, or X0.0767 = 
NH,. PtXO.5453 -NHiCl, or X0.1839 -NH*, or X0.1736 =NH,. 



VOLUMETRIC METHODS FOR DETERMINATION OF 

AMMONIA 

Two conditions are considered: 

A, Estimation of free ammonia in solution. 

B, Determination of ammonia in its salts — combined ammonia. 

Analysis of Aqua Ammonia 

Provided no other basic constituent is present, free ammonia in solution 
is best determined by direct titration with an acid in presence of methyl orange 
or methyl red as indicator. 

Procedure. About 10 grams of the solution in a weighing bottle with glass 
stopper is introduced into an 800-cc. Erlenmeyer flask containing about 200 cc. 
of water and sufficient J normal sulphuric acid to combine with the ammonia 
and about 10 cc. in excess. The flask is stoppered and warmed gently. This 
forces out the stopper in the weighing bottle, the ammonia combining with 
the acid. Upon thorough mixing, the solution is cooled, and the excess of acid 
is titrated with half normal caustic. 

One cc. i N. H2SO4 =0.0085 gram NH,. 
Factor. H2SO4 X 0.3473 =NH,. 

Note. The aqua ammonia exposed to the air will lose ammonia, hence the sample 
should be kept stoppered. This loss of ammonia is quite appreciable in strong 
ammoniacal solutions. 

Determination of Combined Ammonia. Ammonium Salts. 

Strong bases decompose ammonium salts, liberating ammonia. This may be 
distilled into standard acid or into a saturated solution of boric acid (neutral to 
methyl orange) and titrated. 

Procedure. About 1 gram of the substance is placed in a distillation flask 
(see Fig. 50) and excess of sodium or potassium hydroxide added and the 
ammonia distilled into a saturated solution of boric a^id or an excess of standard 
sulphuric acid. Ammonia in boric acid solution may be titrated directly with 
standard acid (methyl orange or methyl red indicator) or in case a mineral acid 

* Factors recpmmended by Tread well and Hall, " Analytical Chemistry," 2, John 
Wiley & Sons. ~ 



298 NITROGEN 

was used to absorb the ammonia, the excess of acid is titrated with standard 
caustic solution. 

One cc. half normal sulphuric acid =0.0085 gram NHi. 
One cc. normal acid =0.01703 gram NH,. 
Factors. HiSO* X 0.3473 =NH, and NH, X 2.8792 =HjS04. 



ANALYSIS OF AMMONIACAL LIQUOR 

The crude liquor by-product from coal gas in addition to ammonia contains 
hydrogen sulphide, carbon dioxide, hydrochloric acid, sulphuric acid, combined 
with ammonia, also sulphites, thiosulphates, thiocyanates, cyanides, ferrocyanides, 
phenols. 

Determination of Ammonia 

Volatile Ammonia. This is determined by distillation of the ammonia into 
an excess of standard sulphuric acid and titrating the excess of acid. With the 
exception that caustic soda is omitted in this determination, the details are the 
same as those for total ammonia as stated in the next paragraph. 

Total Ammonia. The true value of the liquor is ascertained by its total 
ammonia content. Ten to 25 cc. of the sample is diluted to about 250 cc. in 
a distilling flask with a potash connecting bulb, as previously described, 20 cc. 
of 5% sodium hydroxide are added and about 150 cc. of solution distilled into an 
excess of sulphuric acid. The excess is then titrated according to the standard 
procedure for ammonia. 

One cc. N. H2SO4 =0.01703 gram NH,. 
Fixed Ammonia is the difference between the total and the volatile ammonia. 

Carbon Dioxide 

Ten cc. of the liquor are diluted to 400 cc. and 10 cc. of 10% ammoniacal 
calcium chloride added and the mixture, placed in a flask with Bunsen valve, is 
digested on the water bath for two hours. The precipitated calcium carbonate 
is washed, placed in a flask and an excess of N/2 HCl added and the excess acid 
titrated with N/2 NaOH. 

N/2 HCl =0.011 gram CO,. 

Hydrochloric Acid 

Ten cc. of the liquor is diluted to 150 cc. and boiled to remove ammonia. 
Now hydrogen peroxide is added to oxidize organic matter, etc., the mixture being 
boiled to remove the excess of the peroxide. Chlorine is titrated in presence of 
potassium chromate as indicator by tenth normal silver nitrate after neutralize 
ing with dilute nitric acid. 

One cc. N/10 AgNO, =0.00364 gram HCl. 



NITROGEN 299 



Hydrogen Sulphide 

To 10 cc. of the liquor are added an excess of ammoniacal zinc chloride or 
acetate, the mixture diluted to about 80 cc. and warmed to 40®. After settling 
for half an hour the zinc sulphide is filtered off and washed with warm water 
(40 to 50®); the precipitate is washed from the filter into an excess of N/10 
iodine solution, the sr^phide clinging to the paper washed into the main solu- 
tion with hydrochloric acid. The mixture is acidified and the excess iodine 
titrated with N/10 sodium thiosulphate. 

One cc. N/10 1 =0.0017 gram H,S or 0.0016 gram S. 

Sulphuric Acid 

250 cc. of the liquor is concentrated to 10 cc, 2 cc. of concentrated hydro- 
chloric added and the mixture heated to decompose any thiosulphate, sulphide 
or sulphite present. The concentrate is extracted with water, filtered and made 
to 250 cc. The sulphuric acid is now precipitated in an aliquot portion with 
barium chloride. 

BaSO4X0.4202 =H,S04, or X0.1374 =S present as H2SO4. 

Total Sulphur. Fifty cc. of the liquor is run by means of a pipette into 
a deep beaker (250 cc. capacity), containing an excess of bromine covered by 
dilute hydrochloric acid. The mixture is evaporated to dryness on the steam 
bath and the residue taken up with water and diluted to 250 cc. Sulphur is 
now precipitated as barium sulphate as usual, preferably on an aliquot portion. 

For a more complete analysis of crude liquor determining sulphite, thio- 
sulphate, thiocyanate, hydrocyanic acid, ferrocyanic acid, and phenols the analyst 
is referred to Lunge, " Technical Methods of Chemical Analysis," Part II, Vol. II, 
D. Van Nostrand Co. 

Determination of Traces of Ammonia 

The determination of traces of ammonia is best accomplished by the colori- 
metric method with Nessler's reagent. Details of the procedure are given in the 
chapter on water analysis. 



NITRIC ACID. NITRATES 

The alkalimetric method for determining free nitric acid, and the complete 
analysis of the conunercial product are given in the chapter on Acids. Special 
procedures for determining the combined acid are herein given. 

Qravimetric Method for Determining Nitric Acid by Precipitation 

as Nitron Nitrate, C20H10 N4HNO3 

As in case of ammonia the volumetric methods are generally preferable for 
determining nitric acid, combined or free. Isolation of nitric acid by precipita- 
tion as nitron nitrate may occasionally be used. The fairly insoluble, crystalline 
compound, C2oHi6N4«HN08 is formed by addition of the base diphenyl-endo- 



300 NITROGEN 

anilo-hydro-triazole (nitron) to the solution containing the nitrate as directed 
under Separations. The precipitate washed with ice-water is dried to constant 
weight at 110** C. 16.53% of the compound is NOi. 

Note. The following acids should not be present in the solution, since their 
nitron salts are not readily soluble: nitrous, chromic, chloric, perchloric, hydrobromici 
hydroiodic^ hydroferrocyanic, hydroferricyanic, oxalic, picric and thiocyanic acids. 

Solubihty of less soluble nitron salts m 100 cc. of water. Nitron nitrate— 0.0099 
gn^, nitron bromide =0.61 gram, iodide =0.017 gram, nitrite =0.19 gram, chromate 
=0.06 gram, chlorate 0.12 gram, Jperchlorates 0.008 ^ram. thiocyanate=0.04 gram. 
(Treadwell and Hall, ^'Analytical Chemistry, Quantitative Analysis.") 



VOLUMETRIC METHODS 

Direct Estimation of Nitrates by Reduction to Ammonia. 

Modified Devarda Method ^ 

An accurate procedure for the determination of nitrogen in nitrates is Allen's 
modification of the Devarda method. The method is based upon the quantita- 
tive reduction of nitrates to anunonia in an alkaline solution by an alloy con- 
sisting of 45 parts of aluminum, 50 parts of copper and 5 parts of zinc. The 
ammonia evolved is distilled into standard sulphuric acid and thus estimated. 
The method, originally designed for the valuation of sodium or potassium nitrates, 
is also of value in the determination of nitric acid, nitrites or ammonia. In the 
latter case the alloy is omitted. 

Reagents Required. Devarda's Alloy. Forty-five parts aluminum, 50 parts 
copper and 5 parts zinc. The aluminum is heated in a Hessian crucible in a fur- 
nace -until the aluminum begins to melt, copper is now added in small portions 
until liquefied and zinc now plunged into the molten mass. The mix is heated 
for a few moments, covered and then stirred with an iron rod, allowed to cool 
slowly with the cover on and the crystallized mass pulverized. 

Standard Sulphuric Acid, This is made from the stock C.P. acid by dilu- 
tion so that 1 cc. is equal to 0.0057 gram H2SO4, 100 cc. of acid of this strength 
being equivalent to approximately 1 gram of sodium nitrate. (A tenth normal 
acid will do, a smaller sample being taken for analysis.) Since it is necessary to 
standardize this acid against a standard nitrate, it is advisable to have an acid 
especially for this determination rather than a conunon reagent for general use. 

Standardization of the Acid, 11.6 grams of standard potassium nitrate, 
equivalent to about 9.6 grams of NaNOs, is dissolved and made to volume in the 
weighing bottle (100 cc), and 10 cc. is placed in the Devarda flask, reduced and 
the ammonia distilled into 100 cc. of the acid, exactly as the following method 
describes. The temperature of the acid is noted and its value in terms of H2SO4, 
KNOj and NaNOi stated on the container. The acid expands or contracts 
0.029 cc. for every degree centigrade above or below the temperature of stand- 
ardization. 

Standard Potassium Nitrate, The purest nitrato that can be obtained is 
recrystallized in small crystals, by stirring, during the cooling of the supersatu- 
rated concentrated solution, and dried first at 100° C. for several hours and then 

* Paper by W. S. Allen, General Chemical Company, Eighth International Con- 
gress of Applied Chemistry. 



NITROGEN 



301 



at 210® C. to constant weight. Chlorides, sulphates, carbonates, lime, magnesia 
and sodium are tested for and if present are determined and allowance made. 

Standard Sodium Hydroxide. This should be made of such strength that 1 cc. 
is equal to 1 cc. of the standard acid, 2 cc. methyl red being used as indicator. 
Ten cc. of the acid are diluted to 500 cc. and the alkali added until the color of 
the indicator changes from a red to a straw color. 

Methyl Red Solution. 0.25 gram of methyl red is dissolved in 2000 cc. of 
95% alcohol; 2 cc. of the indicator is used for each titration. As the indicator 
is sensitive to COj, all water used must first be boiled to expel carbonic acid. 
(Baker & Adamson, manufacturers of methyl red.) 

Sodium Hydroxide — Sp.gr. 1.3. Pure sodium hydroxide is dissolved in 
distilled water and boiled in an uncovered casserole with about 1 gram of 
Devarda's alloy to remove ammonia. This is cooled and kept in a well-stoppered 
bottle. 

Apparatus. This is shown in the accompanying illustration, Fig. 51 . It consists 
of the Devarda flask connected to the scrubber A', filled with glass wool. This 



Absorpfhn Has/a 



ITSce. 
Crienmeyer If 
Flask 

2ccHiSQ^ 



=\ 




300cx,j 
EHenmeyeri 
Flask 

H2SO4 I 
Scrubber! ^ 




275 cc. 
■Devarda Flask 



^ (5 y'"5''Cas5erofe 

•Icf Sample 
25 cc. of 20XK0H 



■Stand with 
Windshield 



"Meker Burner 



w y yy/////////////^^^^ 



Fig. 51. — Devarda's Apparatus. 

scrubber is heated by an electric coil or by steam passed into the surrounding 
jacket. The scrubber prevents caustic spray from being carried over into the 
receiving flask 0. The form of the apparatus can best be ascertained from the 
sketch. 

Weighing bottle with graduation at 100 cc. and a 10-cc. dropper with rubber 
bulb is used for weighing out the sample in solution. See Fig. 52. 



302 NITROGEN 

Preparation of the Sample 

Weight. It is advisable to take a large sample if possible, e.g., 100 grams 

of NaNOa, 119 grams of KNO, or about 80 grams of strong HNOi (95%) or more 

/^^ if the acid is dilute. Solids are taken from a large sample, all 

( " i 1 fllS^ lumps being broken down. After dissolving in water the sample 

is made up to 1 liter. (Scum is broken up by addition of a 

little alcohol.) One hundred cc. of this solution is placed in 

the weighing bottle, which has been previously weighed, being 

perfectly clean and dry. The difference is the weight of the 

100-cc. sample. 

Manipulation. All parts of the apparatus are washed out 
<»«« with COt-free water. All water used in this determination 
should be boiled to expel COj. Ninety-eight cc. of the standard 
^^ acid is placed in flask and washed down with 2 to 3 cc. of 
"1 water. Two cc. of the standard acid is placed in flask P and 
V"^" y washed down with 10 cc. of water and 13 to 14 drops of methyl 
Fig. 52. red indicator added. Connections are made between the flasks 
WeiKhine Bottle ^^^ ^^® scrubber. (The correction is made for the acid, the 
and Dropper. ^"^P^rat^re being noted at the time of withdrawal.) A cas- 
serole, filled with cold water, is placed under F (see illustration). 
The stem E is removed from the Devarda flask and 10 cc. (or more) of the 
nitrate added by means of the dropper in the weighing bottle, a funnel having been 
inserted in the flask. The bottle reweighed gives the weight of the sample 
removed, by difference. The nitrate is washed down with 10 cc. of water and 
25 cc. of 20% caustic added (free from NH,), the alkali washed down with 10 cc. 
more of water and then 3 grams of Devarda alloy placed in the flask by means of 
dry funnel. The stem E is quickly replaced, the stopcock being turned to close 
the tube. The reaction begins very soon. If it becomes violent, the reaction 
may be abated by stirring the water in the casserole, thus cooling the sample. 
After the energetic action has abated (five minutes), the casserole with the cold 
water is removed and the action allowed to continue for twenty minutes, mean- 
time heat or steam is turned on in the scrubber. E is connected at C to the 
flask B containing caustic to act as a scrubber. It Ls advisable to have a second 
flask containing sulphuric acid attached to the caustic to prevent ammonia from 
the laboratory entering the system. A casserole with hot water is placed under 
F and the burner lighted and turned on full. A gentle suction is now applied 
at Rj the stop-cock D being turned to admit pure air into the evolution flask ; the 
rate should be about 5 to 6 bubbles per second. The suction is continued for 
thirty minutes, hot water being replaced in the casserole as the water evaporates. 
The heat is now turned off and the apparatus disconnected at M and J, The 
contents of this elbow and the condenser are washed into the flask 0. The acid 
in and P poured into an 800-cc. beaker and rinsed out several times. The 
volume in the beaker is made up to 500 cc, 1 cc. of methyl red added, and the 
free acid titrated with the standard caustic. The end-point is a straw yellow. 
Calculation. The cc. of the back titration with caustic being deducted, 
the volume of the acid remaining (e.g., combined with ammonia) is corrected to 
the standard condition. Expansion or contraction of the acid is 0.029 cc. per each 
degree C. above or below the temperature at which the acid was standardized. 
If the acid is exactly 0.057 gram H1SO4 per cc, the result multipUed by 0.989 and 



NITROGEN 303 

divided by the weight of the sample taken gives per cent nitrate. (In terms of 
NaNO,.) 

The Weight of the Sample. Ten times the difference of the weighings 
of the bottle W before and after removal of the 10 cc. and the product divided 
by the weight of the 100 cc. of the solution equals the weight of solid taken. 

Example. Weight of the bottle -\- 100 cc. sample = 218 grams. Weight of the bottle 
= 112 ^rams, therefore weight of 100 cc. NaNOi = 106 grams. 

Weight oi the bottle -h 100 cc. sample =218. Weight after removal of 10 cc= 207.4 
grams, therefore sample taken = 10.6 grams, including the added water. Now from 
above the weight of tne actual sample taken = 10.6 X 10 -r 106 = 1 gram. 

Temperature Correction, Temperature of standardization =20** C. Temperature 
of the sulphuric acid when taken for the analysis =31** C. Back titration of the 
caustic = 2 cc. The correct volume = (100 - 2) -((31- 20) X 0.029) =97.681 cc. 
HaSOi combined with ammonia from the reduced nitrate. 97.681 X0.989 -^ 1 = 96.62% 
NaNO,. 

Factors. H2SO4 X 2.06107 =KNOa or Xl.7334=NaNO, or Xl.2850=HNO,. 

HtSO4X0.9587 = HNO2 or X0.3473=NH,. 

NHaX3.6995=HNO, or X4.9906 = NaNO, or X 4.0513 =NaNO,. 

NaNOaX 1.1894 = KNO, and KNO,X 0.8408 = NaNO,. 



ANALYSIS OF NITRATE OF SODA 

The following impurities may occur in nitrate of soda: KNO», NaCl, Na2S04, 
NaaCO,, NaClOa, NaClO*, Fe^O,, AUO,, CaO, MgO, SiO,, H,0, etc. In the 
analysis of sodium nitrate for determination of NaNOa by difference, moisture, 
NaCl, Na2S04 and insoluble matter are determined and their sum deducted 
from 100, the difference being taken as NaNOa. Such a procedure is far from 
accurate, the only reliable method being a direct determination of niter by the 
Devarda method given in detail. The following analysis may be required in the 
valuation of the nitrate of soda. 

Determination of Moisture 

Twenty grams of sample are heated in a weighed platinum dish at 205 to 
210** C. for fifteen minutes in an air bath or electric oven. The loss of weight 
multiplied by 5 = per cent moisture. (Save sample for further tests.) 

Insoluble Matter 

Ten grams are treated with 50 cc. of water and filtered throu^ a tared 
Gooch. The increased weight dried residue (100° C.) multiplied by 10= per 
cent insoluble matter. (Save filtrate.) 

Sodium Sulphate 

The moisture sample is dissolved in 20 cc. hot water and transferred to a 
porcelain crucible. It is evaporated several times with hydrochloric acid to 
dryness to expel nitric acid. (Until no odor of free chlorine is noticed when thus 
treated.) Fifty cc. of water and 5 cc. hydrochloric acid are now added and the 



304 NITROGEN 

sample filtered. Any residue remaining is principally silica. The filtrate is 
heated to boiling, 10 cc. of 10% barium chloride solution added, and the precipi- 
tated sulphate filtered ofif, ignited and weighed. 

BaSOiX 3.0445 =per cent NajSO*. 

Iron, Alumina, Lime, and Magnesia 

These impurities may be determined on a 20-gram dried sample, the 
material being dried and evaporated as in case of the sodium sulphate determina- 
tion. The filtrate from silica is treated with ammonium hydroxide and Fe(OH)i 
and A1(0H)» filtered ofif. Lime is precipitated from the iron and alumina filtrate 
as oxalate and magnesia determined by precipitation as phosphate from the 
lime filtrate by the standard procedures. 

Sodium Chloride 

The filtrate from the insoluble residue is brought to boiling and magnesia, 
MgO (CI free), is added until the solution is alkaline to litmus. 0.5 cc. of 1% 
potassium chromate (KzCrOO solution is added as an indicator and then the 
solution is titrated with a standard solution of silver nitrate until a faint red 
tinge is seen, the procedure being similar to the determination of chlorides in 
water by silver nitrate titration. The cc. AgNOa X factor for this reagent X 10 = per 
cent NaCl. 

Silver nitrate is standardized against a salt solution. 

Carbonates 

This determination is seldom made. CO2 may be tested for by addition of 
dilute sulphuric acid to the salt. Efifervescence indicates carbonates. Any 
evolved gas may be tested by lime water, which becomes cloudy if CO2 is present. 
For details of the procedure reference is made to the chapter on Carbon. 



DETERMINATION OF NITRIC NITROGEN IN SOIL EXTRACTS 

Vamari-Mitscherlich-Devarda Method 

Procedure. Forty cc. of water, a small pinch of magnesia and one of mag- 
nesium sulphate are added to flask D of the Mitscherlich apparatus (Fig. 53). 
Twenty-five cc. of standard acid and GO cc. of neutral redistilled water are placed 
in flask F; 250 or 300 cc. of aqueous soil extract are placed in a 500-cc. Kjeldahl 
flask, 2 cc. of 50% sodium hydroxide added, the mouth of the flask closed with 
a small funnel to prevent spattering, and the contents of the flask boiled for 
thirty minutes. The water which has boiled off is replaced, and, after cooling, 
1 gram of Devarda's alloy (GO mesh), and a small piece of paraffin are added 
and the flask connected with the apparatus; reduction and distillation arc carried 
on for forty minutes. The receiver contents are then cooled, 4 drops of 0.02% 



NITROGEN 



305 



solution of methyl red added^ the excess acid is nearly neutralized, the liquid 
boiled to expel COa, cooled to 10 to 15® and the titration completed. 




Hopkins 
Tube 



Z 




500 c.c. 
Jena , 
Kjeldahl 
Flask 



ZOOc.c. 
Jena 
Flask 




^Heavy Walled 
Quartz Tube 
4.5mm. I. D, 

E 



y 



Zcm. ^7(r/77. 



ZOOcc 
Erlenmeyer 
Flask 



yi///'///////////////////7>}}}}}}??//m 




Fig. 53. — Mitscherlich's Apparatus for Nitrogen Determination. 



DETERMINATION OF NITROGEN OF NITRATES (AND 
NITRITES) BY MEANS OF THE NITROMETER 

The nitrometer is an exceedingly useful instrument employed in the accurate 
measurement of gases liberated in a great many reactions and has therefore a 
number of practical applications. It may be used in the determination of carbon 
dioxide in carbonates; the available oxygen in hydrogen dioxide; in the valua- 
tion of nitrous ether and nitrites; in the valuation of nitrates and nitric acid 
in mixed acids. 

The method for the determination of nitrogen in nitrates, with which we 
are concerned in this chapter, depends on the reaction between sulphuric acid 
and nitrates in presence of mercury: 

2KNOa+4H2S04+3Hg=K2S04+3HgS04+4H,0+2NO. 

The simplest type of apparatus is shown in the illustration. Fig. 54. The 
graduated decomposition tube has a capacity of 100 cc. It is connected at the 
base by means of a heavy-walled rubber tubing with an ungraduated leveling 
tube (6). At the upper portion of (a) and separated from it by a glass stopn 
cock (s) is a bulb (c) of about 5 cc. capacity; a second stop-cock enables com- 
pletely enclosing the sample, as may be necessary in volatile compounds. The 
glass stop-cock (s), directly above the graduated chamber, is perforated so as 



306 NITROGEN 

to establish connection with the tube (d) when desired and the graduated 
cylinder (a). 

Procedure. The tube (6) is filled with mercury and the air in (a) now dis- 
placed by mercury, by turning the stop-cock to form an open passage between 
(a) and (d) and then raising (6). A sample of not over 0.35 gram potassium 
nitrate or a corresponding amount of other nitrates, is introduced into (c), the 
material being washed in with the least amount of water necessary (1 to 2 cc). 
By lowering (6) and opening the stop-cock s the solution is drawn into the 
decomposition chamber, taking care that no air enters. This is followed by about 
15 cc. of pure, strong sulphuric acid through ^i and s, avoiding admitting air as 
before. NO gas is liberated by the heat of reaction between the sulphuric acid 
and the water solution. When the reaction subsides, the tube (a) is shaken 
to mix the mercury with the liquor and the NO completely liberated. The gas 
is allowed to cool to room temperature and then measured, after raising or 
lowering (b) so that the column of mercury is the calculated excess of height 
above that in (a) in order to have the gas under atmospheric pressure. The 
excess of height is obtained by dividing the length of the acid layer in (a), in 
millimeters, by 7 and elevating the level of the mercury in (6) above that in (a) 
by this quotient; i.e., if the acid layer =21 mm. the mercury in (b) would be 
3 mm. above that in (a). The volume of gas is reduced to standard conditions 
by using the formula 

V(P-w) 

760(1+0.003670' 

F'= volume under standard conditions; F= observed volume; P= observed 
barometric pressure in mm.; m;= tension of aqueous vapor at the observed tem- 
perature, expressed in millimeters; i = observed temperature. 

One cc. gas =4.62 milligrams of KNOj, or 3.8 milligrams NaNOi 

or 2.816 milligrams HNO3. 

Du Pont Nitrometer Method * 

The Du Pont nitrometer. Fig. 55, is the most accurate apparatus for the volu- 
metric determination of nitrates. By use of this, direct readings in per cent may 
be obtained, without recourse to correction of the volume of gas to standard con- 
ditions and calculations such as are required with the ordinary nitrometers. 

The apparatus consists of a generating bulb of 300 cc. capacity E with its 
reservoir F connected to it by a heavy-walled rubber tubing. E carries two 
glass stop-cocks as is shown in illustration. The upper is a two-way stop-cock 
connecting either the cup or an exit tube with the chamber. D is the chaml>er- 
reading burette, calibrated to read in percentages of nitrogen, and graduated 
from 10 to 14%, divided in 1/100%. Between 171.8 and 240.4 cc. of gas must 
be generated to obtain a reading. A is also a measuring burette, that may be 
used in place of D where a wider range of measurement is desired. " It is used 
for the measurement of small as well as large amounts of gas. It is most com- 
monly graduated to hold 300.1 milligrams of NO at 20° C. and 760 mm. pres.surc 
and this volume is divided into 100 units (subdivided into tenths) each unit 
being equivalent to 3.001 milligrams of NO. When compensated, the gas from 

^See paper by J. R. Pitman, Jour. See. Chem. Ind., p. 983, 1900. 



NITROGEN 



307 



ten times the molecular weight in milligrams of any nitrate of the formula RXOi 
(or five timea molecular weight of R(NOi)i) should exactly fill the burette. This 
simplifies all calculations; for example the per cent nitric acid in a mixed acid 
would be 

-j^^ = percentHNO.. 

fl— burette reading, H'=granis acid taken."' C is the compensating burette 
very similar in form to the chamber burette Z). B ia the leveling bulb, by the 






FiQ- M. — Nitrometer. 



Fio. 55. — Du Poat's Nitrometer. 



raisiDg or lowering of which the standard pressure in the system nmy be obtained. 
The apparatus as shown in Fig. 55 is mounted on an iron stand. As in the more 
simple form of apparatus, previously described, mercury is used as the con- 
fining liquid. The parts are connected by heavy-walled rubber tubing, wired 
to the glass parts. 

> A. W. Belts, Chemist, E. I. DuPont de Nemours Powder Co., \a letter to author. 



308 NITROGEN 

Standardizing the Apparatus. The apparatus having been arranged and the 
various parts filled with mercury, the instrument is standardized as follows: 
20 to 30 cc. of sulphuric acid are drawn into the generating bulb through the 
cup at the top, and at the same time about 210 cc. of air; the cocks are* then 
closed, and the bulb well shaken; this thoroughly desiccates the air, which is then 
run over into the compensating burette until the mercury is about on a level with 
the 12.30% mark on the other burette, the two being held in the same relative 
position, after which the compensating burette is sealed off at the top. A 
further quantity of air is desiccated in the same manner and run into the read- 
ing burette so as to fill up to about the same mark; the cocks are then closed, 
and a small piece of glass tubing bent in the form of a U, half filled with sul- 
phuric acid (not water), is attached to the outlet of the reading burette; when 
the mercury columns are balanced and the enclosed air cooled down, the cock 
is again carefully opened, and when the sulphuric balances in the U-tube, and 
the mercury columns in both burettes are at the same level, then the air in each 
one is under the same conditions of temperature and pressure. A reading is now 
made from the burette, and the barometric pressure and temperature carefully 
noted, using the formula 

_ FoPo(273-f 
' Pi273 ' 

the volume this enclosed air would occupy at 29.92 ins. pressure and 20° C. is 
found. The cock is again closed and the reservoir manipulated so as to bring 
the mercury in both burettes to the same level, and in the reading burette to the 
calculated value as well. A strip of paper is now pasted on the compensating 
burette at the level of the mercury, and the standardization is then complete. 

Another rapid method of standardizing is to fill the compensating chamber 
with desiccated air as stated in the first procedure and then to introduce into the 
generating chamber 1 gram of pure potassium nitrate dissolved in 2 to 4 cc. of 
water, the cup is rinsed out with 20 cc. of 66** B6aume sulphuric acid, making 
three or four washings of it, each lot being drawn down separately into the bulb. 
The generated gas formed after vigorous shaking of the mixture, as stated under 
procedure, is run into the measuring burette. The columns in both burettes are 
balanced so that the reading burette is at 13.85 (=per cent N in KNOj). A 
strip of paper is pasted on the compensating burette at the level of the mer- 
cury, and standardization is accomplished. By this method the temperature 
and pressure readings, and the calculations are avoided.^ 

Procedure for Making the Test. Salts, One gram of sodium or potassium 
nitrate, or such an amount of the material as. will generate between 172 to 240 cc. 
of gas, is dissolved in a little water and placed in the cup of the generating bulb. 

Liquid Acids. The acid is weighed in a Lunge pipette and the desired amount 
run into the funnel of the generating bulb, the amount of acid that is taken 
being governed by its nitrogen content. 

The sample is drawn into the bulb; the funnel is then rinsed out with three 
or four successive washings of 95% sulphuric acid, the total quantity being 

20 cc. 

To generate the gas, the bulb is shaken well until apparently all the gas is 

^ Standardization with " C. P. KNOi is the better, as it is less tedious and is not 
subject to the correction errors that cannot be escaped when standardizing with air. 
The KNOj must be of undoubted purity." — ^A. W. Betts. 



NITROGEN 309 

formed, taking care that the lower stop-cock has been left open, this cock is then 
closed and the shaking repeated for two minutes. The reservoir is then lowered 
until about 60 cc. of mercury and 20 cc. of acid are left in the generating bulb. 
There will remain then sufficient space for 220 cc. of gas. 

Note. If too much mercury is left in the bulb, the mixture will be so thick that 
it will be found difficult to complete the reaction, a long time will be required for the 
residue to settle and some of the gas is liable to be held in suspension by the mercury, 
so that inaccurate results follow. 

The generated gas is now transferred to the reading burette, and after wait- 
ing a couple of minutes to allow for cooling, both burettes are balanced, so that 
in the compensating tube the mercury column is on a level with the paper mark 
as well as with the column in the reading burette; the reading is then taken. 

If exactly one gram of the substance is taken the percentage of nitrogen may 
be read directly, but in case of other amounts being taken, as will invariably be 
the case in the analysis of acids, the readings are divided by the weight of the 
substance and multiplied by 4.5 to obtain the per cent of nitric acid mono- 
hydrate present. 

The procedure may be used for determining nitrites as well as nitrates. 

Determination of HNO3 in Oleum by Du Pont Nitrometer Method * 

About 10 cc. oleum are weighed in a 30-cc. weighing bottle, 10 cc. 95% 
reagent sulphuric acid added and mixed by shaking. This mixture is transferred 
to the nitrometer reaction tube and the weighing bottle and nitrometer cup 
rinsed with three 5-cc. portions of the reagent sulphuric acid which is drawn into 
the reaction tube. This is vigorously shaken for three minutes and the gas 
then passed to the measuring tube and allowed to stand for about five minutes, 
after which the mercury levels are adjusted and the reading taken. 

It is obvious that this determination includes any nitrous acid in the oleum. 

Combined Nitric Acid 

The nitric acid in nitrates may be determined by titration with ferrous 
sulphate. The nitrate dissolved in a little water is run into strong sulphuric acid 
and titrated with standard ferrous sulphate according to the procedure described 
for determining free nitric acid in mixed acids in the chapter on the subject. 

* By courtesy of £. I. du Pont de Nemours Powder Co. 



PHOSPHORUS 

Wilfred W. Scott 

P4, at. wt. 31.04; sp.ffr. < ^^ 2.296* '^'^' 1 725'' * '^^ 1 • oxuiea, 

P,0„ PO„ P,0-, acids, HJPO2, HJPO,, HJPO4, HPO„ H4P1O7. 

DETECTION 

Element Phosphorus is recognized by its glowing (phosphorescence) in 
the air. The element is quickly oxidized to P2O5; if the yellow modification is 
slightly warm (34** C.) the oxidation takes place with such energy that the 
substance bursts into flame. The red form is more stable. It ignites at 260° C. 

Boiled with KOH or NaOH it forms phosphine, PHi, which in presence of 
accompanying impurities is inflammable in the air. 

Phosphorus oxidized to P1O5 may be detected with ammonium molybdate, 
a yellow compound, (NH4)8P04'12Mo03*3HiO, being formed. 

Acids. Hypophosphorous Acid, HiPOt, heated with copper sulphate to 
55**. C. gives a reddish-black compound, CujHi, which breaks down at 100° to 
H and Cu. Permanganates are reduced immediately by hypophosphorous acid. 
No precipitates are formed with barium, strontium or calcium solutions. Zinc 
in presence of sulphuric acid reduces hypophosphorous acid to phosphine, PH3. 

Phosphorous Acid, HiPOi. Copper sulphate is reduced to metallic copper 
and hydrogen is evolved, no CujHj being formed as in case of hypophosphorous 
acid. Permanganates are reduced slowly. Added to solutions of barium, stron- 
tium or calcium white phosphites of these elements are precipitated. Alkali 
phosphites are soluble in water, while hypophosphites are not readily soluble. 

Orthophosphoric Acid, HsPO^. Ammonium phosphomolybdate precipi- 
tates yellow ammonium phosphomolybdate from slightly nitric acid solutions. 
The precipitate is soluble in ammonium hydroxide. 

Metaphosphoric Acid, HPOz. Converted by nitric acid in hot solutions 
to the ortho form. Metaphosphoric acid is not precipitated by ammonium 
molybdate. 

Pyrophosphoric Acid, HiP^O^, Converted to orthophosphoric acid in hot 
solutions by nitric acid. No precipitate is formed with anunonium molybdate. 

Comparison of Ortho, Meta and Pyrophosphoric Acids 



Reagent. 



Orthophosphoric 
acid. 



Ammonium molybdate 

Albumin 

Zinc sulphate, cold, in excess. . . . 
Silver nitrate in neutral solution. 

Magnesium salts 



Yellow ppt. 



Yellow ppt., 
. Ag,P( )4 
White ppt. 

310 



Metaphosphorio 
acid. 



No ppt. 
Coagulated 
No ppt. 
White ppt., 

AgPO, 
No ppt. 



Pyrophoaphorio 
acicf. 



No ppt. 
Not coagulated 
White ppt. 
White ppt., 
Ag^hOr 
No ppt. 



PHOSPHORUS 



311 



Phosphorous acids are distinguished from phosphoric acids by the phosphine 
formed with the former when acted upon with zinc. 

Acid phosphates are distinguished from normal phosphates as follows: 
Neutral silver nitrate added to an acid phosphate liberates free nitric acid (Litmus 
test), the following reaction taking place: 

^3AgNO,+NaaHP04=Ag,P04+2NaN6,+HNO,. 

The solution resulting when silver nitrate is added to normal phosphate 
solution is neutral. 

3AgN0,+Na JO4 = AgJ>04+3NaNO,. 



ESTIMATION 

The determination of the pentoxide of phosphorus is required in a large 
number of substances, since it is widely distributed in the form of phosphates — 
calcium phosphate, Ca3(P04)i; fluor apatite, 3Ca3(P04)2'CaFi; chlor apatite, 
3Ca,(P04),-CaCl2; vivianite, Fe,(P04)2-8H20;wavelite, 2Al2(P04)fAl2(OH)6-9H20; 
pyromorphite, 3Pb3(P04)2'PbCl2; phosphates of iron and calcium in phosphate 
ores, hence in slags of the blast furnace. It occurs in fertile soils, bones, plant 
and animal tissues. 

The chemist is especially concerned in the determination of phosphoric 
acid (P2O6), in the evaluation of materials used for the manufacture of the acid — 
bone ash and phosphate rock (see table below). Generally, determinations 
of lime, iron and alumina are also desired and frequently a more complete 
analysis. In the analysis of phosphoric acid certain impurities occurring in 
the crude material used are determined, e.g., iron, lime, magnesia, sulphuric, 
hydrochloric and hydrofluoric acids, etc. Phosphoric acid is determined in 
the evaluation of phosphate fertilizers, phosphates used in medicine, phosphate 
baking powders, etc. 

The element is determined in iron, steel, phosphor bronzes, and other alloys. 

Typical Analyses* 



Subatance. 



Bone Ash. 



Phosphoric oxide 
Sulphur trioxide. 
Carbon dioxide. . 

Lime.. 

Magnesia 

Alumina 

Ferric oxide 

Fluorine, etc 

Alkaline salts . . . 
Silica — sand, etc. 



39.55 


52.46 
1.02 


6!i7 


oisi 



Charlestown 


Spanish 


Sombrero 


Redonda 


Phosphate. 


Pho-^phorite. 


Phosphate. 


Phosphates. 


27.17 


33.38 


35.12 


35.47 


3.30 


0.57 






4.96 


4.10 


7.40 




44.03 


47.16 


51.33 




0.37 


trace 






1.44 


0.89 


+Fe 


20.17 


0.43 


2.59 


1.02 


8.85 


2.38 


4.01 






0.87 




0.42 




5.60 


3.71 


2.02 


9.70 



Canadian 
Phosphate. 

37.68 



51.04 

FeiOa, 

Al,0„ 

F. etc. 

= 6.88 

4.29 



* Thorpe, " Dictionary of Applied Chemistry," Longmans, Green & Co. 

Preliminary Remarks. Practically all procedures for the determination of 
phosphorus depend upon iU oxidation to ortho phosphoric acid and its pre- 
cipitation by ammonium molybdate from a nitric acid solution as ammonium 
phoepho-molybdate. It may now be determined either gravimetrically or 
volumetrically. Two procedures are of importance in the gravimetric deter- 



312 PHOSPHORUS 

mination of phosphorus; the first depends upon the direct weighing of the 
yellow phosphomolybdate, dried at 110° C; the second, on the conversion of the 
yellow precipitate to the magnesium salt and its ignition to pyrophosphate. 
Two volumetric procedures, which are of special value in the determination of 
small amounts of phosphorus as in case of phosphorus in iron and steel, are to be 
recommended for their rapidity and accuracy. One of these is to dissolve the 
ammonium phosphomolybdate in a known amount of standard caustic, titrate 
the excess of alkali with standard acid, which indicates the alkali required to 
neutralize the molybdic acid in the yellow precipitate. From this the 
amount of phosphorus present may be calculated. A second procedure of equal 
accuracy and rapidity is to dissolve the molybdate in ammonia, add an excess 
of sulphuric acid, pass the warm solution through a column of zinc and titrate 
the reduced molybdic acid with standard potassium permanganate, the amoimt 
of permanganate required being a measure of the phosphorus present. 

The impurities interfering in the procedures are silica and arsenic acid. The 

first may be eliminated by dehydration of the silicic acid in the solution and its 

removal as insoluble SiO* by filtration. Arsenic in small quantities does not 

interfere under certain conditions; in large quantities its removal is imperative. 

Preparation and Solution of the Sample 

Amount of the Sample Required. For accurate results it is advisable to 
take a fairly large sample, 5 to 10 grams, and when it has been dissolved, to 
dilute to a definite volume, 500 or 1000 cc. Aliquots of this solution are ttJcen 
for analysis. 

Iron Ores, Phosphate Rock and Minerals. Five to 10 grams of the pul- 
verized material placed in a 3-in. porcelain dish are digested for an hour with 
50 to 100 cc. of concentrated hydrochloric acid (sp.gr. 1.19), the dish being covered 
by a clock-glass and placed on a steam bath. The acid is now diluted with 
half its volume of water and the solution filtered into a porcelain dish of suf- 
ficient capacity to hold the filtrate and washings. The residue is washed with 
dilute hydrochloric acid (1:1) until free of visible iron discoloration. The 
filtrate and washings are evaporated rapidly on a hot plate to small volume 
and then to dr)mess over the steam bath. Meanwhile the insoluble residue and 
filter are ignited in a 20-cc. platinum crucible over a M^ker burner or in a muffle 
furnace and the residue fused with ten times its weight of sodium carbonate. 
The fusion is removed by inserting a platinum wire into the molten mass, 
allowing to cool and then gently heating until the mass loosens from the cru- 
cible, when it may be removed on the wire. The cooled mass on the wire and that 
remaining in the crucible are dissolved in dilute hydrochloric acid, and the 
filtered solution added to the main solution. The combined solutions are 
evaporated to dryness, and heated gently to dehydrate the silica. The residue 
is taken up with a few cc. of hydrochloric acid, the solution diluted, filtered 
and the SiOx washed with dilute nitric acid solution. The combined filtrates 
are made up to 500 or 1000 cc. Aliquots of this solution are taken for analysis. 

Iron and Steel. Five to 10 grams of the drillings or filings are dissolved in an 
Erlenmeyer flask with 50 to 100 cc. of dilute nitric acid, 1:1, more acid being 
added if necessary. When dissolved, a strong solution of KMn04 is added 
until a pink color appears; on boiling brown manganese dioxide forms in the 
solution if a sufficient amount of permanganate has been added. This is dis- 



PHOSPHORUS 313 

solved by adding 2% sodium thiosulphate solution in just sufficient quantity to 
dissolve the precipitate. The solution is diluted to a convenient volume for 
analysis. Where a number of determinations are to be made, it is advisable 
to weigh the amount of sample desired for the determination and to precipitate 
the ammonium phosphomolybdate in the flask in which the drillings have be^n 
dissolved. 

Ores Containing Titanium. Titanium may be recognized by the red color 
produced by hydrogen peroxide, HjOj, added to the sulphuric acid extract; also 
by the reduction test with zinc, which causes a play of colors, the solution becom- 
ing colorless by the reduction of iron, then, in presence of titanium, pink, purple 
and finally blue. (Vanadium gives similar tests.) Solutions containing titanium 
frequently appear milky when the solution is diluted before filtering off the insol- 
uble residue. Since titanium forms an insoluble compound with phosphoric acid 
and iron oxide ^ the final residue, obtained by the method of solution for ores, 
phosphate rock and minerals, should be moistened with sulphuric acid and the 
silica expelled with hydrofluoric acid. The solution is evaporated to dryness and 
to S0» fumes, the residue fused with sodium carbonate and taken up with boiling 
water. TiOi remains insoluble, while P2O.J passes into the filtrate as the sodium 
salt. The procedure may be shortened by treating the original sample directly 
according to this method of solution, a 2-gram sample being taken, as larger 
amounts are difficult to handle. 

Soluble Phosphates, Phosphate Baking Powder, etc. A water extract 
is generally sufficient to get the material in solution. In case iron, alumina, lime 
and magnesia salts are present, as may occur in baking powders, an extraction 
with dilute 3% nitric acid is necessary. It is advisable to dissolve a 5- to 10-gram 
sample and take an aliquot part of the solution made up to a definite volume. 
Before precipitatmg with ammonium phosphomolybdate, 5 grams of ammonium 
nitrate should be added for each gram of the sample taken for analysis and the 
solution boiled to oxidize compounds of phosphorus to the orthophosphate 
form. 

Precipitation of Ammonium Phosphomolybdate 

Precipitation of ammonium phosphomolybdate is common to all subsequent 
methods for determination of phosphorus^ andy as in case of preparation and solu- 
tion of the sample, details of this procedure wiU not be repeated. 

Reaction. 

H J>04+12(NH4),Mo04+21HNO, = (NH4) JPO* • 12MoO,+21NH4NO,+12H,0. 

Special Reagents Required. Ammonium Molyhdate. One hundred grams 
of pure molybdic acid are thoroughly mixed with 400 cc. of cold distilled water 
and 80 cc. of strong ammonia (sp.gr. 0.90) added. When the solution is com- 
plete it is poured slowly and with constant stirring into a mixture of 400 cc. 
of strong nitric acid (sp.gr. 1.42) and 600 cc. distilled water. This order of 
procedure should be followed, as the nitric acid poured into the ammonium 
molybdate solution will cause the precipitation of a difficultly soluble oxide of 
molybdenum and render the reagent practically worthless. Fifty milligrams 
(0.50 gram) of micrccosmic salt, dissolved in a little water are added, the pre- 
cipitate agitated, then allowed to settle for twenty-four hours and the clear solu- 

^ Blair " Chemical Analysis of Iron.'' 



314 PHOSPHORUS 

tion decanted through a filter into a large reagent bottle. Sixty cc. of the reagent 
should be used for every 0.1 gram of PjOs present in the solution analyzed. 

Potassium Permanganate. For oxidation purposes. Two per cent solution 
filtered free of dioxide through asbestos is required. 

Amount of Sample Required for Analysis. If the material contains over 
20% P2O6, 0.1 to 0.5-gram sample should be taken; if the product contains 
5 to 20% PiOs, 1.0 to 0.5 gram should be taken; for a sample containing 0.5 
to 5%, 2.5 to 1-gram sample is taken, and for PjOs less than 0.5%, a 5-gram sam- 
ple is taken. 

Precipitation. The free acid of the solution is nearly neutralized by addi- 
tion of ammonium hydroxide. In analysis of phosphate rock or materials com- 
paratively low in iron, it is advisable to add ammonium hydroxide in quantity 
sufficient to cause a slight permanent precipitate followed by just sufficient HNOi 
to dissolve the precipitate. In iron and steel analysis ammonium hydroxide is 
added until the precipitated iron hydroxide dissolves with difficulty and the 
solution becomes a deep amber color or cherry red. In analysis of soluble phos-' 
phateSf litmus paper dropped into the solution indicates the neutral point. 
Nitric acid is added to the neutral or slightly acid solution, 5 cc. of acid for every 
100 cc. of solution. A volume of 150 to 200 cc. of solution is the proper dilu- 
tion for samples taken in amounts above recommended. To the warm solu- 
tion (not over 80** C.) ammonium molybdate is added, 60 cc. of the reagent being 
required for every 0.1 gram of PjOs present. The solution is stirred, or shaken, 
if in a flask, until a cloudy precipitate of anmionium phosphomolybdate appears. 
It is then allowed to settle on the steam bath at a temperature of 40 to 60** C, 
for one hour, then again agitated and allowed to settle in the cold for an hour 
longer. The filtrate should be tested with additional ammonium molybdate for 
phosphorus. The yellow precipitate is filtered and washed with 1% HNO« solu- 
tion followed by a 1% solution of KNOa, or NH4NO3 or (NH4)2S04 as the special 
case requires. Filtration through asbestos in a Gooch crucible is to be recom- 
mended. When a large number of determinations are to be made, as in case of 
iron and steel, filter paper is more convenient. 



QRAVIMETRIC METHODS FOR DETERMINATION OF 

PHOSPHORUS 

A. Direct Weighing of the Ammonium Phosphomolybdate 

The sample being dissolved and the ammonium phosphomolybdate precipitated 
according to directions already given above, the supematont solution is filtered 
through a weighed Gooch crucible and washed twice by decantation with 
dilute nitric acid (1%), the precipitate washed into the Gooch, followed by two 
washings with 1% KNOi or NH4NO3 (neutral solutions) and finally with water. 
The precipitate, free from contaminating impurities, is dried for two hours in an 
oven at 110° C, then cooled in a desiccator and weighed. Weight of precipi- 
tate X 0.0 165= P, or X 0.03784 =P,06. 

Note. If this procedure is to be followed it will be convenient to take 1.65 grams 
sample, if the phosphorus content will allow. Each 0.01 gram of precipitate will then 
equal 1 % P. 



PHOSPHORUS 315 



B. Determination of Phosphorus as Magnesium Pyrophosphate 

Magnesia Mixture. For precipitation of ammonium magnesium phos- 
phate, 110 grams of magnesium chloride (MgCU«6H20) are dissolved in a small 
amount of water. To this are added 280 grams of ammonium chloride and 
700 cc. of ammonia (sp.gr. 0.90); the solution is now diluted to 2000 cc. with 
distilled water. The solution is allowed to stand several hours and then filtered 
into a large bottle with glass stopper. Ten cc. of the solution should be used for 
every 0.1 gram P2O6 present in the sample analyzed. As the reagent becomes 
old it will be necessary to filter off the silica that it gradually accumulates from 
the reagent bottle. 

Procedure. The ammonium phosphomolybdate, obtained as directed (page 
314), is filtered onto a 12J S. & S. No. 689 filter paper and washed four or five times 
with dilute 1% HNOi. The precipitate is now dissolved from the filter by a fine 
stream of hot ammonium hydroxide, 1:1, catching the solution in the beaker in 
which the precipitation was made. The solution and washings should be not over 
100 to 150 cc. Hydrochloric acid is added to the cooled solution to neutralize the 
excess of ammonia, the yellow precipitate, that forms during the neutralization, 
dissolving with diflSculty, when sufficient acid has been added. To the cooled 
solution cold magnesia mixture is added drop by drop (2 drops per second) with 
constant stirring. Ten cc. of the reagent will precipitate 0.1 gram PjOs. When the 
solution becomes cloudy the stirring is discontinued and the precipitate allowed 
to settle ten minutes. Ammonium hydroxide is added until the solution con- 
tains about one-fourth its original volume of strong ammonia (e.g. 25 cc. NH4OH, 
90 to 100 cc. of solution). The solution is stirred during the addition and then 
allowed to settle for at least two hours. It is filtered preferably, thi:ough a 
Gooch crucible (or through an ashless filter paper), and the precipitate washed 
with dilute ammonium hydroxide, 1 : 4, then placed in a porcelain crucible, a 
few drops of saturated solution of ammonium nitrate added and the precipitate 
heated over a low flame till decomposed (or until the paper chars). The lumps 
of residue are broken up with a platinum rod and again ignited over a Scimatico 
or M6ker burner, the heat being gradually increased. If the heating is properly 
conducted, the resultant ash will be white or light gray, otherwise it will be 
dark. The addition of solid ammonium nitrate aids the oxidation in obstinate 
cases, but there is danger of slight mechanical loss. The crucible is cooled in a 
desiccator and the residue weighed as magnesium pyrophosphate. 

MgiPaOr X 0.2787 =P and MgJ^jOjX 0.6379= PA. 

Direct Precipitation of Magnesium Ammonium Phosphate 

In the absence of heavy metals whose phosphates are insoluble in an ammo- 
niacal solution, the magnesia mixture may be added directly to the neutral solu- 
tion containing the phosphate, without previous precipitation of ammonium 
phosphomolybdate. The magnesium ammonium phosphate is washed and 
ignited according to directions given above, and weighed as magnesium pyro- 
phosphate. 

The use of the Gooch crucible for the ammonium phosphomolybdate and the 
ammonium magnesium phosphate precipitates is recommended in preference 
to filter paper, as the fibers of the latter invariably are occluded in the precipi- 



316 PHOSPHORUS 

tates, and produce dark-colored residues of magnesium pyrophosphate, which are, 
frequently, extremely difficult to bum white. The residue on the asbestos mat, 
on the other hand, is easily ignited white, and does not require repeated 
addition of an oxidizing agent, as is so often the case with precipitates filled 
with paper fiber. 



VOLUMETRIC METHODS FOR THE DETERMINATION 

OF PHOSPHORUS 

C. Alkalimetric Method 

The method is based, on the acid character of ammonium phosphomolybdate, 
the following reaction taking place with an alkali hydroxide: 

2(NH4),12MoOJ»04+46NaOH+H,0 

= 2(NH4)2HP04+ (NH4),Mo04+23Na2Mo04+23H,0. 

From the reaction 46 molecules of sodium hydroxide are equivalent to one 
molecule of PjOs, hence 1 cc. of a N/10 solution of sodium hydroxide neutralizes 
the yellow precipitate containing an equivalent of .000309 gram of PjOf. 

Special Reagents 

Sodium Hydroxide — Tenth Normal Solution. For determination of phos- 
phorus by the alkali volumetric method. To 100 grams of the pure NaOH sticks 
sufficient water is added to just dissolve the hydrate. This concentrated solu- 
tion is poured into a tall cylinder, the vessel closed and the insoluble matter allowed 
to settle. The liquid will be practically free of carbonates. A portion of the clear 
liquor may now be drawn off and diluted to a definite volume so that the solu- 
tion is slightly stronger than tenth normal, as determined by titration against 
a standard N/10 acid. It may now be diluted to the required amount as indi- 
cated by the acid titration. Freshly boiled distilled water should be used in the 
dilutions of the standard caustic solution. Phenolphthalein indicator is required 
in the titration. The exact value of the caustic solution in terms of phosphorus 
may be ascertained by standardizing the solution against a steel sample of known 
phosphorus content, the sample being dissolved in nitric acid, the phosphorus 
precipitated as ammonium phosphomolybdate, the washed precipitate dissolved 
by the caustic solution and the excess caustic titrated by standard nitric acid 
according to the procedure given later. 

Wt. of P in sample . i. t^ i. xt ^tx 

■ ^, ^^--. : — y- — . ■ , , = amount of P per cc. of NaOH. 

cc. of NaOH required to neutralize inolybdate 

Nitric Acid — Tenth Normal Solution. The acid is standardized against 
the caustic solution and should be of such 'strength that 1 cc. of HNOs is equal 
to 1 cc. of NaOH. Phenolphthalein indicator is used. Approximately 6.7 cc. 
of 95% HNO, diluted to 1000 cc. =N/10 HNO, solution. 

Nitric Acid for Washing Precipitates. One per cent solution, 14 cc. HNOi 
(sp.gr. 1.42) per liter of water. 

Acid Ammonium Sulphate for Washing Precipitates. Fifteen cc. strong 
NH4OH+25 cone. H,S04 in 1000 cc. solution. 



PHOSPHORUS 317 

Potasaium Nitrate for Washing Precipitates, Used in volumetric analjrsis 
only. Ten grams of KNOi per liter of solution. Test, to be sure the solution 
is neutral. 

Other reagents required: NH4OH (sp.gr. 0.90); HjSO* (sp.gr. 1.84); HNOi 
(sp.gr. 1.42); NasSsOt solution, 2%; amalgamated zinc. 

Special Apparatus Required 

Jones* Reductor. Details of the reductor are given under the determination 
of iron by the permanganate method, also under the Volumetric Determination 
of Molybdenum, pages 220 and 281. 

Procedure. The ammonium phosphomolybdate, obtained according to direc- 
tions already given on page 313, is filtered into a Gooch crucible containing 
asbestos, and washed once or twice with water containing 1% nitric acid, and then 
several times with a 1% neutral solution of potassium nitrate until the washings 
are free of acid, as indicated by testing with litmus paper. The asbestos mat con- 
taining the precipitate is transferred to a No. 4 beaker, 100 cc. of COj free water 
added, followed by about 20 cc. of N/10 NaOH measured from a burette. The 
crucible is rinsed out with 5 to 10 cc. of N/10 NaOH, the exact amount being 
noted and then with water, adding the rinsings to the main solution. Phenol- 
phthalein indicator is added, and the excess of caustic titrated with N/10 HNO«. 
The total NaOH added minus the acid titration equals the cc. of the caustic 
equired to react with the yellow precipitate. 

One cc. N/10 NaOH =0.000136 gram P and =0.000309 gram PjOs. 

The exact factor should be determined as directed under Reagents. 

D. Zinc Reduction and Titration with Potassium Permanganate 

This method is based od the assumption that ammonium phosphomolybdate, 
(NH4)3l2MoOiP04, is reduced, in acid solution, by zinc, the molybdic acid, MoOi, 
forming the lower oxide MojOi, in which form it reacts with ferric iron in the 
receiving flask, reducing a corresponding equivalent of ferric salt to ferrous 
condition, being itself oxidized to MoOi.* When the ferric solution is not placed 
in the receiving flask a slight oxidation takes place, the oxide MouOn, apparently 
being formed.' 

Potassium Permanganate Decinormal Solution. For volumetric determina- 
tion of phosphorus, reduction method, 3.161 grams of the pure salt per liter is 
the theoretical amount required for a tenth-normal solution. It is necessary, 
however, to standardize the permanganate solution against a tenth-normal 
sodium oxalate solution. The exact value of the permanganate solution may be 
accurately and rapidly determined in terms of phosphorus by standardizing against 
a sample of standard steel containing a known amount of phosphorus, the 
ultimate standard being steel drillings furnished by the U. S. Bureau of Standards. 
The drillings are dissolved in nitric acid, oxidized with KMn04, the excess of the 
reagent being destroyed by thiosulphate solution. Ammonia is added until the 
solution becomes a deep amber color. The phosphorus is precipitated as 
ammoniimi phosphomolybdate. The following procedure is the same as is 

» D. L. Randall, Am. Jour. Sci. (4), 24, 316. 

« Blair, " Chemical Analysis of Iron," 7th Ed., p. 96. 



318 PHOSPHORUS 

given in the volumetric method following: The permanganate titration of the 

reduced molybdic acid divided into the amount of phosphorus known to be 

present in the solution will give the value of the permanganate in terms of 

phosphorus. 

Wt. of P in sample ^ i. t> /. t^h r ^ 

^^^, ,^ r-— 7 = amount of P per cc. of KMnO*. 

cc. KMnOi required 

Procedure. The ammonium phosphomolybdate, obtained by the procedure 
given on page 313, is filtered onto an asbestos mat in a Gooch crucible or onto 
filter paper, and washed with dilute HNO3 followed by acid ammonium sulphate 
(15 cc. NH4OH, sp.gr. 0.90+25 cc. cone. H^SO*, sp.gr. 1.84+1000 cc. H^O), until 
2 or 3 cc. of the wash water gives no reaction for molybdenum with a drop of 
ammonium sulphide solution. Five or six washings should suffice. 

Reduction, The precipitate is dissolved by adding about 10 cc. ammonium 
hydroxide, sp.gr. 0.96, to the precipitate, catching the solution in the beaker or 
flask in which the precipitation was made. About 10 cc. of strong sulphuric 
acid is added to this solution after diluting to about 100 cc. The Jones reductor 
is prepared as described for determination of molybdenum by reduction, page 281. 
The receiving flask is charged with about 25 cc. of ferric alum (100 grams per liter) 
and 4 cc. syrupy phosphoric acid. In iron and steel analysis this mixture is omitted. 
One hundred cc. of hot water followed by 100 cc. of hot dilute sulphuric acid 
(2.50%) are passed through the column of zinc in the reductor (previously cleaned 
by passing hot dilute H2SO4 through it). The phosphomolybdic solution is now 
poured through the reductor, followed by 100 cc. 2% sulphuric acid and 100 cc. 
hot water. The solution as it is reduced becomes green, but upon coming in con- 
tact with the ferric solution produces a bright red color. In absence of ferric solu- 
tion in the receiver the reduced solution appears green and should remain clear. 
The hot solution is titrated immediately with N/10 KMn04. 

Titration. The reduced solution is poured into a No. 6 beaker and N/10 
KMn04 added from a burette, until a faint permanent pink color is produced. 
During the titration, the solution changes in color to a brown, a pinkish yellow 
and finally to pink. 

Titration of reduced ferric solution, 1 cc. N/10 KMnO* =0.000862 P. 

Titration in absence of ferric sulphate, 1 cc. N/10 KMnO* =0.000887 P. 

Calculation. Case 1. Ferric sulphate in the receiver (6MojOj+ 18 O = I2M0O3 
in the molecule containing IP); 18 are equivalent to 36 H, hence N/10 P 
according to this reaction equals at.wt. P divided by (36X1000) =P for 1 cc. 
N/10 KMnOi =0.000862 g. P. 

Case 2. No ferric salt in receiver. Moj40j7+35 0=24MoOi+2P. 
(35 0=70 H). Dividing by 2 we get at.wt. P divided by (35X1000) =P for 
1 cc. N/10 KMnOi =0.000887. 



PHOSPHORUS 319 



Report of the Committee on Research and Analytical Methods — 

Phosphate Rock ^ 

The following tentative standard methods for sampling and deteni ination of 
moisture, phosphoric acid and iron and alumina in phosphate rock are recommended 
to the Division. 

Methods of Sampling and Determination of Moisture 

I. Gross Sample. A. Car Shipments. One hundred pounds sample per car. 

1. Samplina from the Car. In sampling car shipments in the car at least ten 
scoopshovelsfuly aggregating 100 lbs., shall be taken from each car at approximately 
equal distances from each other so as to average the car. Care shall be taken to 
see that each scoopful shall cover the entire face of the pile from floor to top. 

2. Sampling from the Cart or Barrow. A small hand scoopful of 1 to 2 lbs. shall 
be taken from each cart or barrow either as it is being loaded or as it leaves the car. 

B. Cargo Shipments. One hundred pounds minimum sample per vessel. 

1. Sampling in Hoisting Tub. In sampHng cargoes generalljr running from 1000 
tons upward a small hancf scoopful shall be taken from approximately every tenth 
tub before it is hoisted from the hold. 

2. Sampling from Conveyor. If unloading is being done with automatic bucket 
and conveyor, periodical sectioas ol the entire discharge of the conveyor shall be taken 
of such intervals and quantity as to give a sample equivalent to approximately 1 lb. 
per each 10 tons of cargo. 

3. Sampling from Conveying Vehicle. Samples shall be taken with a hand scoop 
from various cars at such regular intervals and in such quantities as to give approx- 
imately 1 lb. for each 10 tons of cargo. 

n. Laboratory Sample. The resulting gross sample obtained by any one of the 
methods outUned shall be crushed to pass a four-mesh screen, thoroughly mixed on a 
clean, hard surface and quartered down to a 10-lb. average sample. 

A. Crushing. This 10-lb. sample shall all be crushed to pass an eight-mesh 
screen. 

B. Mixing and Quartering. This eight-mcah sample shall be carefully mixed 
and quartered down to two 2-lb. samples. 

C. Grinding. 1. Moisture Sample. One of these 2-lb. samples shall be held in an 
air-tight contamer. This sample is to be used for the determination of moisture. 

2. Analytical Sample. The other 2-lb. sample shall be further mixed and quartered 
down to a 2- or 4-oz. sample which is then to be ground to pass a sixty-mesh screen 
or preferably a sixty-five mesh screen. This sample is to be used for the analytical 
determination. 

Note. It is essential that the taking of the gross sample be done with small 
hand scoops and that the practice of taking the sample in the hand be absolutely 
prohibited, for it has been found that there is considerable selective action in the 
finer materials sifting through the fingers while a scoop retains the entire sample. 

The dimensions of the screens referred to above are to be as follows: 

No. of Mesh Sue of Owning ^ia-ter c,f Wire 

4 0.185 0.065 

8 0.093 0.032 

65 0.0082 0.0072 

m. Determination of Moisture. Moisture is to be determined on both the 
moisture sample and analytical sample. Of the moisture sample not less than 100 
^rams are to be weighed out for each determination. Of the analytical sample approx- 
unately 2 grams are to be weighed out for each determination. Both are to be dried 
to constant weight at a temperature of 105° C. in a well-ventilated oven, preferably 
with a current of dry air passing through the oven. The containers in which moist- 
ure is determined should be provided with well-fitting covers so that the samples 
may be cooled and weighed in the well-covered container. 

^ Journ. Ind. and Eng. Chem. 



320 PHOSPHORUS 

IV. Calculation of Results. The percentages of phosphoric acid and iron and 
alumina as determined on the analytical sample are to be calculated to a moisture- 
free basis and subsequently to the basis of the original sample as shown by the moist- 
ure content of the moisture sample. 

Detennination of Phosphoric Acid 

Reagents, To be prepared as in Official Methods, A. O. A. C. Bureau of Chemistry, 
Bulletin 107 (Rev.)i 1910, p. 2. Preparation of reagents (c), (rf), (c) and (/), except 
that the ammonium nitrate solution m (d) is changed to 5% instead of 10%. 

Method of Solution, To 5 grams of the sample add 30 cc. of cone, hydrochloric 
acid (sp.gr. 1.20) and 10 cc. of cone, nitric acid (sp.gr. 1.42) and boil down to a 
syrupy consistency. The residue, which should be nearly sohd after cooling, is taken 
up with 5 cc. of cone, nitric acid and 50 cc. of water. Heat to boiling, cool, filter 
and make up to 5(X) cc. through the filter. This procedure eliminates practically 
all of the silica and it is necessary to filter as quicKly as possible after digestion so 
as to avoid redissolving the silica. 

Determination. Draw off an aliquot portion of 50 cc, corresponding to 0.5 gram, 
neutralize with ammonia, then add nitric acid imtil the solution is iust clear. Ada 
15 grams of ammonium nitrate (free from phosphates), heat the solution to 50^ C. 
and add 150 cc. of molybdate solution. Digest at 50° C. for fifteen minutes with 
frequent stirring. Filter off the supernatant liquid and test the filtrate with molyb- 
date solution to see if precipitation has been complete. (If not, add more molybdate 
to the filtrate and digest for fifteen minutes longer.) Wash with 5 per cent ammoniuin 
nitrate solution by decantation, retaining as much of the precipitate as possible in the 
beaker. Dissolve the precipitate in the beaker in the least possible quantity of ammo- 
nium hydroxide (sp.gr. 0.90) and dilute this solution with several times its volume 
of hot water. Dissolve the remainder of the precipitate on the filter with this solu- 
tion, washing beaker and filter with hot water and keeping the volume of the filtrate 
between 75 and 100 cc. Neutrahze with hydrochloric acid, cool to room tempera- 
tuie and add 25 cc. of magnesia mixture from a burette, drop by drop, stirring 
vigorouslv with a rubber-tipped rod, then add 15 cc. of ammonium hydroxide (sp.gr. 
0.90) ana allow to stand for four hours or overnight at room temperature. The time 
of standing may be reduced to two hours if kept in a refrigerator or still better in an 
ice-water bath. Filter through a platinum or porcelain Gooch crucible, fitted with 
a platinum or asbestos mat carefully made and ignited to constant weight. Wash 
with 2.5% ammonium hydroxide until practically free from chlorides; dry, ignite, 
cool and weigh as magnesium pyrophosphate. If desired, filtration may be made 
through an aSiless filter paper, igniting m the usual manner. Calculate to PsOi by 
multiplying by 0.6378 (log 80468). 

Determination of Iron and Aluminum together as Phosphates 

1. Solutions Required: 1. Hydrochloric acid (1 : 1); prepared by mixing 1 part 
by volume of concentrated HCl (sp.gr. 1.19) with 1 part of distilled water. 

2. A saturated solution of ammonium chloride, which should be filtered before 
use. 

3. A 25% solution of ammonium acetate, faintly acid to litmus paper. 

4. A solution of ammonium phosphate (10%). prepared by dissolving 20 grams of 
(NH4)iHP04 in 180 cc. of distilled water and filtering. (This should be prepared 
frequently in small Quantity, as it attacks glass containers on standing.) 

5. A standard solution of ferrous ammonium sulphate, containing iron equivalent 
to about 0.0100 gram of FojOi in 10 cc. and 50 cc. cone. HCl per liter. 

6. A solution of calcium and magnesium phosphates for blank determinations, 
prepared as follows: Dissolve 4 grams of MgO and 35 grams of CaCOj (both free 
of iron and aluminum) in 100 cc. concentrated HCl, add an aqueous solution of 30 
grams of (NH4)2HP04, make up to 2 liters and -filter. 

7. A solution of ammonium nitrate (5%) for washing precipitates. About 400 cc. 
are required for each determination. 

All reagents used should be as pure as practicable and all solutions should be free 
of suspended matter. 

n. Preparation of Rock Solution. Place 2.5 grams of pulverized rock with 50 cc. 
of 1 : 1 HCl in a graduated 250-cc. flask, the glass of which contains less than 1% 



PHOSPHORUS 321 

of iron and aluminum oxides.^ Boil ^ntlv with occasional shaking for one hour in 
such a manner as to avoid concentratmg the solution to less than half of its original 
volume,* dilute, cool to room temperature, make up to volume and mix; niter 
immediately throu^ a dry filter into a dry nask, discarding the first few co. of the 
filtered solution. 

Pipette a 50-cc. aliquot, representing 0.5 gram of rock,